{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"github-narcooo--inkos","slug":"narcooo--inkos","name":"inkos","type":"agent","url":"https://github.com/Narcooo/inkos","page_url":"https://unfragile.ai/narcooo--inkos","categories":["automation","code-review-security"],"tags":["agent","ai","ai-agent","ai-novel","ai-writing","chinese-novel","cli","cli-tool","creative-writing-ai","novel-ai","novel-generation","openclaw"],"pricing":{"model":"open_source","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"github-narcooo--inkos__cap_0","uri":"capability://planning.reasoning.multi.agent.orchestrated.novel.generation.pipeline","name":"multi-agent orchestrated novel generation pipeline","description":"Coordinates a specialized 10-agent pipeline (Radar, Planner, Composer, Architect, Writer, Continuity Auditor, Reviser, and others) where each agent handles specific creative and logical tasks in sequence. Agents communicate through a shared Truth Files state system that maintains canonical world state, character matrices, and plot hooks across the entire generation process. The pipeline enforces human review gates between critical stages, allowing writers to approve or reject agent outputs before proceeding to the next phase.","intents":["I want to automate the end-to-end novel writing process while maintaining narrative consistency across 100+ chapters","I need agents to specialize in different aspects (planning, writing, auditing, revision) rather than one monolithic LLM","I want to insert human review checkpoints so I can approve plot directions before agents write entire chapters"],"best_for":["fiction writers and indie authors scaling novel production","publishing teams managing multiple concurrent novel projects","creators building serialized content (web novels, light novels) with strict continuity requirements"],"limitations":["Pipeline latency scales with chapter length and number of agents — typical full cycle (plan→write→audit→revise) takes 2-5 minutes per chapter","Human review gates create blocking dependencies; if a human rejects output, the entire downstream pipeline must restart","Agent coordination overhead adds ~15-20% computational cost vs single-model generation due to inter-agent communication and state synchronization","Requires careful tuning of agent prompts and Truth Files schema per genre; generic configuration produces mediocre results"],"requires":["Node.js 18+","TypeScript 5.0+","API keys for at least one LLM provider (OpenAI, Anthropic, Ollama, or compatible)","Minimum 2GB RAM for local state management and agent coordination"],"input_types":["text (story premise, genre, character descriptions)","JSON/YAML (configuration files, genre profiles)","markdown (Truth Files, story bible, writing rules)"],"output_types":["text (generated chapters, revision notes)","JSON (state snapshots, audit reports)","markdown (Truth Files updates, continuity logs)"],"categories":["planning-reasoning","automation-workflow","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_1","uri":"capability://memory.knowledge.truth.files.state.management.system","name":"truth files state management system","description":"Maintains 7 canonical markdown and JSON files that serve as the persistent knowledge base for the entire novel generation pipeline: World Bible (setting/lore), Character Matrix (detailed character profiles with relationships), Plot Hooks (story beats and foreshadowing), Chapter Snapshots (summaries of each completed chapter), Writing Rules (genre-specific and universal constraints), and Continuity Log (detected inconsistencies and resolutions). Agents read from and write to these files via a State Manager API that enforces schema validation using Zod and TypeBox, preventing malformed state updates.","intents":["I need a single source of truth for character details so agents don't contradict established facts across chapters","I want to track plot threads and foreshadowing so the story remains coherent over 200+ chapters","I need to audit what state changed between chapters to detect and fix continuity errors automatically"],"best_for":["long-form fiction projects (novels, web serials, light novels) requiring strict continuity","collaborative writing teams where multiple agents must agree on canonical facts","writers who want to version-control and diff their story state across revisions"],"limitations":["Truth Files are stored as local markdown/JSON files; no built-in cloud sync or collaborative locking — multiple agents writing simultaneously can cause race conditions","Schema changes require manual migration of existing Truth Files; no automatic schema evolution","File I/O latency adds ~50-100ms per agent read/write cycle; high-frequency state updates (e.g., per-sentence auditing) become bottlenecked","Truth Files grow linearly with chapter count; a 500-chapter novel produces ~2-5MB of state files, requiring careful indexing for fast lookups"],"requires":["Node.js 18+","File system write access (local or mounted network drive)","Zod 3.20+ and TypeBox 0.28+ for schema validation"],"input_types":["markdown (World Bible, Character Matrix, Writing Rules)","JSON (Chapter Snapshots, Continuity Log, Plot Hooks)","text (agent-generated updates to state)"],"output_types":["markdown (updated Truth Files)","JSON (state diffs, validation reports)","text (continuity audit reports)"],"categories":["memory-knowledge","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_10","uri":"capability://text.generation.language.fanfic.and.spinoff.mode.with.source.material.integration","name":"fanfic and spinoff mode with source material integration","description":"Allows users to provide existing source material (published novels, fanfic, or original works) as reference context, enabling agents to generate spinoffs, prequels, or alternate universe stories that maintain consistency with the source. The system extracts key characters, plot points, and world details from the source material and populates the Truth Files automatically, reducing manual setup. Agents are instructed to respect source canon while exploring new story directions.","intents":["I want to write a prequel to an existing novel while maintaining consistency with the original story","I want to generate a fanfic that respects the source material's characters and world-building","I want to create an alternate universe version of a story with different plot outcomes"],"best_for":["fanfic writers who want AI assistance while respecting source canon","authors writing prequels or spinoffs of their own works","publishers exploring derivative works based on existing IP"],"limitations":["Automatic extraction of characters and plot points from source material is imperfect; manual review and correction required","Agents may contradict source canon despite instructions; human review is essential","Copyright concerns: generating fanfic of copyrighted material may have legal implications depending on jurisdiction","Source material must be provided as plain text; PDFs, ePubs, or other formats require manual conversion"],"requires":["Node.js 18+","Source material as plain text or markdown","Clear understanding of source canon and desired spinoff direction"],"input_types":["text (source material: existing novel, fanfic, or original work)","text (spinoff premise: prequel, alternate universe, etc.)"],"output_types":["JSON (extracted characters, plot points, world details)","text (generated spinoff chapters)","JSON (canon consistency audit report)"],"categories":["text-generation-language","memory-knowledge"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_11","uri":"capability://planning.reasoning.architect.agent.for.plot.outlining.and.story.structure","name":"architect agent for plot outlining and story structure","description":"A specialized agent that generates detailed plot outlines and story structures based on a high-level premise. The Architect reads the genre profile and writing rules, then produces a hierarchical outline (acts, chapters, scenes) with plot beats, character arcs, and foreshadowing notes. The output is stored in Truth Files (Plot Hooks) and used by downstream agents (Planner, Writer) to maintain narrative coherence. The Architect can be invoked standalone to generate outlines for human review before any writing begins.","intents":["I want to generate a detailed plot outline before agents start writing chapters","I need a structured story arc that respects genre conventions and writing rules","I want to see the full story structure (acts, chapters, character arcs) before committing to generation"],"best_for":["writers who want to plan the entire story before generation begins","projects with complex plots (multiple POVs, subplots) that require careful structure","teams that need to approve the story outline before agents write chapters"],"limitations":["Outline generation is time-consuming; a 100-chapter outline takes 5-10 minutes to generate","Outlines are suggestions; agents may deviate from the outline during writing, requiring manual correction","Complex plots with multiple subplots can produce confusing or contradictory outlines","Outline changes require manual updates to Truth Files; no automatic propagation to downstream chapters"],"requires":["Node.js 18+","High-level story premise (1-2 paragraphs)","Genre profile and writing rules defined"],"input_types":["text (story premise, genre, target length)","YAML/JSON (genre profile, writing rules)"],"output_types":["markdown (hierarchical plot outline with acts, chapters, scenes)","JSON (structured outline with metadata: character arcs, foreshadowing, plot beats)"],"categories":["planning-reasoning","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_12","uri":"capability://planning.reasoning.planner.and.composer.agents.for.chapter.planning.and.structure","name":"planner and composer agents for chapter planning and structure","description":"The Planner agent breaks down the overall plot outline into individual chapter plans, specifying which plot beats, characters, and scenes should appear in each chapter. The Composer agent then structures the chapter plan into a detailed scene-by-scene breakdown with dialogue notes, pacing guidance, and emotional beats. Both agents read from Truth Files (Plot Hooks, Character Matrix) and write their outputs back to Truth Files for the Writer agent to consume. This two-stage planning ensures chapters are coherent and aligned with the overall story structure.","intents":["I want agents to plan each chapter before writing it, ensuring it fits the overall story structure","I need detailed scene breakdowns so the Writer agent knows exactly what to write","I want to review chapter plans before agents write the full text"],"best_for":["complex stories with multiple POVs or subplots that require careful chapter planning","writers who want to approve chapter structure before writing begins","projects where chapter consistency is critical (e.g., serialized web novels)"],"limitations":["Planning adds 1-2 minutes per chapter before writing begins; total generation time increases by 30-50%","Detailed plans can constrain Writer agent creativity; overly prescriptive plans produce formulaic chapters","Plans may become outdated if earlier chapters deviate from the outline; manual re-planning required","Scene-by-scene breakdowns are difficult to generate accurately; complex scenes may be oversimplified"],"requires":["Node.js 18+","Plot outline from Architect agent (or manual outline)","Character Matrix and World Bible in Truth Files"],"input_types":["JSON (plot outline from Architect)","markdown (Truth Files: Character Matrix, World Bible, Plot Hooks)"],"output_types":["JSON (chapter plan with plot beats, characters, scenes)","markdown (scene-by-scene breakdown with dialogue notes and pacing)"],"categories":["planning-reasoning","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_13","uri":"capability://text.generation.language.writer.agent.for.chapter.generation.with.context.awareness","name":"writer agent for chapter generation with context awareness","description":"The core writing agent that generates chapter text based on chapter plans from the Composer agent. The Writer reads from Truth Files (Character Matrix, World Bible, Plot Hooks, Chapter Snapshots) to maintain consistency with established facts and previous chapters. It generates prose in the target style (learned from Style Profile) and respects writing rules (genre-specific and universal). The Writer produces raw chapter text that is then passed to the Auditor for continuity checking and the Reviser for style refinement.","intents":["I want agents to write chapters that are consistent with established characters, world details, and previous events","I need chapters written in a specific style (e.g., Xianxia, Romance) that matches my target audience","I want the Writer to respect writing rules and constraints defined in the genre profile"],"best_for":["any novel generation project using InkOS","writers who want AI assistance with the actual prose writing while maintaining creative control","projects where consistency with established facts is critical"],"limitations":["Writer agent quality depends heavily on chapter plan quality; poor plans produce poor chapters","Context window limitations: for very long novels (500+ chapters), the Writer may lose track of early plot points despite Truth Files","Writing style is learned from Style Profile; insufficient training data (< 10 reference chapters) produces inconsistent style","Writer agent may produce repetitive prose or clichéd dialogue despite style guidance"],"requires":["Node.js 18+","Chapter plan from Composer agent","Truth Files populated with Character Matrix, World Bible, Plot Hooks","Style Profile trained on reference chapters (optional but recommended)","API access to a capable LLM (Claude 3.5 Sonnet or GPT-4 recommended)"],"input_types":["JSON (chapter plan with scenes, dialogue notes, pacing)","markdown (Truth Files: Character Matrix, World Bible, Plot Hooks, Chapter Snapshots)"],"output_types":["text (raw chapter prose, typically 2000-5000 words)","JSON (metadata: word count, character appearances, plot events)"],"categories":["text-generation-language","memory-knowledge"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_14","uri":"capability://safety.moderation.continuity.auditor.agent.for.consistency.checking.and.error.detection","name":"continuity auditor agent for consistency checking and error detection","description":"A specialized agent that audits generated chapters for continuity errors, character inconsistencies, plot contradictions, and AIGC markers. The Auditor reads the chapter text and Truth Files (Character Matrix, World Bible, Chapter Snapshots) and produces a detailed audit report listing detected issues with severity levels (critical, warning, info). Critical issues (e.g., character name misspelling, contradicted plot point) trigger automatic re-generation; warnings (e.g., character behavior inconsistency) are flagged for human review. The Auditor also updates the Continuity Log in Truth Files with detected issues and resolutions.","intents":["I want to automatically detect continuity errors (character name misspellings, contradicted plot points) and fix them","I need to ensure characters behave consistently with their established personalities and backgrounds","I want to track all detected issues and resolutions in a continuity log for future reference"],"best_for":["long-form fiction projects (100+ chapters) where manual continuity checking is infeasible","writers who want automated quality assurance before chapters are published","projects with complex character relationships and plot threads that are easy to contradict"],"limitations":["Auditing is heuristic-based; subtle continuity errors (e.g., character motivation inconsistency) may be missed","False positives are common; the Auditor may flag legitimate plot developments as 'contradictions'","Auditing adds 1-2 minutes per chapter; for a 500-chapter novel, total auditing time is 8-16 hours","Critical issues trigger re-generation, which can loop indefinitely if the Auditor is overly strict"],"requires":["Node.js 18+","Chapter text to audit","Truth Files populated with Character Matrix, World Bible, Chapter Snapshots"],"input_types":["text (chapter to audit)","markdown (Truth Files: Character Matrix, World Bible, Chapter Snapshots)"],"output_types":["JSON (audit report with detected issues, severity levels, and suggested fixes)","markdown (updated Continuity Log in Truth Files)"],"categories":["safety-moderation","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_15","uri":"capability://text.generation.language.reviser.agent.for.style.refinement.and.anti.detect.rewriting","name":"reviser agent for style refinement and anti-detect rewriting","description":"A specialized agent that refines chapter prose to improve readability, remove AIGC markers, and match the target style. The Reviser reads flagged passages from the Auditor (high AIGC probability, style inconsistencies) and rewrites them to be more natural and authentic. It uses the learned Style Profile to guide rewrites and respects writing rules during revision. The Reviser can also perform broader revisions (e.g., 'make this dialogue more natural', 'reduce adjective usage') based on human feedback.","intents":["I want to remove 'AI-tell' markers from generated chapters so they read more naturally","I need to refine prose quality and improve readability without rewriting the entire chapter","I want to match the target author's voice and style in generated text"],"best_for":["writers who want to polish AI-generated prose before publication","projects where authenticity is critical (e.g., publishing to platforms that penalize AI content)","writers with a strong personal style who want AI assistance that matches their voice"],"limitations":["Revision adds 1-2 minutes per chapter; total generation time increases by 20-30%","Over-revision can remove intentional stylistic choices; careful tuning required","Reviser quality depends on Style Profile training; insufficient training data produces poor results","Revision may introduce new errors or inconsistencies; human review recommended"],"requires":["Node.js 18+","Chapter text to revise","Style Profile trained on reference chapters (10-20 samples recommended)","Audit report from Continuity Auditor (optional but recommended)"],"input_types":["text (chapter prose to revise)","JSON (audit report with flagged passages and issues)","markdown (Style Profile and writing rules)"],"output_types":["text (revised chapter prose)","JSON (revision report with changes made)"],"categories":["text-generation-language","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_16","uri":"capability://automation.workflow.radar.agent.for.real.time.monitoring.and.anomaly.detection","name":"radar agent for real-time monitoring and anomaly detection","description":"A background monitoring agent that continuously watches the generation pipeline for anomalies: LLM API failures, Truth Files corruption, generation timeouts, style drift, or suspicious patterns (e.g., repeated character names, plot loops). The Radar agent sends alerts to the human operator and can trigger automatic recovery actions (e.g., retry failed generations, restore Truth Files from backup). It maintains a system health dashboard showing pipeline status, error rates, and performance metrics.","intents":["I want to be alerted immediately if something goes wrong during generation (API failure, timeout, etc.)","I need to monitor the health of the generation pipeline and detect anomalies early","I want automatic recovery actions (retry, rollback) for common failure modes"],"best_for":["long-running generation jobs (100+ chapters) where failures are costly","teams managing multiple concurrent novel projects who need centralized monitoring","production deployments where uptime and reliability are critical"],"limitations":["Radar agent adds ~5-10% computational overhead due to continuous monitoring","Anomaly detection is heuristic-based; false positives (e.g., flagging intentional plot loops as anomalies) are common","Automatic recovery actions may mask underlying issues; human investigation still required","No built-in alerting to external systems (Slack, email, etc.); requires custom webhook integration"],"requires":["Node.js 18+","InkOS core engine running with Radar agent enabled","Optional: webhook URL for external alerts"],"input_types":["JSON (pipeline status, error logs, performance metrics)"],"output_types":["JSON (health dashboard, anomaly alerts)","text (alert messages to human operator)"],"categories":["automation-workflow","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_2","uri":"capability://safety.moderation.aigc.detection.and.anti.detect.revision.loop","name":"aigc detection and anti-detect revision loop","description":"Analyzes generated text using a 33-dimensional auditing system that detects 'AI-tell' markers (repetitive phrasing, unnatural dialogue patterns, overuse of adjectives, structural predictability) characteristic of LLM output. When the Continuity Auditor detects high AIGC probability (>0.6 threshold), it automatically triggers a Reviser agent loop that rewrites flagged passages to increase human-like authenticity. The system maintains a Style Profile per book that learns the target author's voice and uses it to guide revision toward more natural prose.","intents":["I want to detect and remove obvious 'AI-written' markers so my novel doesn't read like ChatGPT output","I need to automatically revise passages that sound too formulaic or repetitive without manual intervention","I want to learn and enforce a consistent author voice across all generated chapters"],"best_for":["indie authors publishing to platforms (Amazon KDP, Webnovel) where AI-generated content is flagged or penalized","writers who want to maintain plausible deniability about AI assistance in their work","publishers managing large volumes of AI-assisted content who need automated quality gates"],"limitations":["33-dimensional auditing is heuristic-based and may produce false positives (flagging intentionally stylized prose as 'AI-like')","Anti-detect revision adds 1-2 minutes per chapter; if many passages are flagged, the revision loop becomes a bottleneck","Style Profile learning requires 10-20 manually-written reference chapters to establish baseline; insufficient training data produces poor style matching","Cannot guarantee detection of sophisticated AI writing that mimics human patterns; adversarial LLMs may evade detection"],"requires":["Node.js 18+","Reference chapters (10-20 human-written samples) to train Style Profile","API access to a capable LLM for revision (Claude 3+ or GPT-4 recommended for nuanced rewriting)"],"input_types":["text (generated chapters to audit)","markdown (reference chapters for Style Profile training)"],"output_types":["JSON (audit report with 33-dimension scores, flagged passages)","text (revised chapters with AI-tell markers removed)","JSON (Style Profile metrics and confidence scores)"],"categories":["safety-moderation","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_3","uri":"capability://tool.use.integration.multi.model.llm.routing.and.cost.optimization","name":"multi-model llm routing and cost optimization","description":"Allows assignment of different LLM providers and models to different agents within the same pipeline, enabling cost-quality tradeoffs. For example, the Writer agent might use Claude 3.5 Sonnet (higher quality, higher cost) while the Continuity Auditor uses GPT-4o (faster, cheaper) and the Reviser uses Ollama locally (free but slower). The system maintains a provider registry that abstracts away API differences, allowing agents to call LLMs through a unified interface regardless of backend (OpenAI, Anthropic, Ollama, or custom providers).","intents":["I want to use expensive models (Claude) only for critical writing tasks and cheaper models (GPT-4o) for auditing to reduce costs","I need to run some agents locally (Ollama) for privacy while using cloud APIs for others","I want to experiment with different model combinations without rewriting agent code"],"best_for":["cost-conscious indie authors generating 100+ chapters who need to optimize per-chapter spend","teams with privacy requirements who want to run some agents locally and others in the cloud","researchers experimenting with different model combinations to find optimal quality/cost ratios"],"limitations":["Provider abstraction adds ~50-100ms latency per LLM call due to request translation and response normalization","Different models produce inconsistent output quality; switching models mid-pipeline can cause downstream agents to receive unexpected formats or styles","Ollama (local) models are significantly slower than cloud APIs; a chapter that takes 2 minutes with Claude takes 10+ minutes with Ollama","No automatic fallback if a provider is down; manual configuration required to switch providers"],"requires":["API keys for at least one LLM provider (OpenAI, Anthropic, or compatible)","Node.js 18+","For Ollama: Ollama runtime installed locally and a model downloaded (e.g., Llama 2, Mistral)"],"input_types":["YAML/JSON (provider configuration, model assignments per agent)","text (prompts sent to LLMs)"],"output_types":["text (LLM responses, normalized across providers)","JSON (cost tracking, latency metrics per provider)"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_4","uri":"capability://automation.workflow.human.in.the.loop.review.gates.with.approval.workflows","name":"human-in-the-loop review gates with approval workflows","description":"Inserts mandatory human review checkpoints between critical pipeline stages (after planning, after writing, before revision) where a human writer must approve, reject, or request modifications to agent output before the pipeline proceeds. Rejected outputs trigger automatic re-generation with modified prompts; approved outputs are locked and passed downstream. The system tracks approval history and allows humans to provide inline feedback that agents incorporate into subsequent generations.","intents":["I want to review the plot outline before agents write the entire chapter based on it","I need to approve major story decisions (character deaths, plot twists) before they're written into the narrative","I want to give agents feedback on their writing and have them revise based on my comments"],"best_for":["professional authors who want AI assistance but maintain creative control","publishing teams with editorial workflows that require approval before content goes live","collaborative projects where multiple humans need to sign off on story decisions"],"limitations":["Human review gates create blocking dependencies; if a human is unavailable, the entire pipeline stalls","Rejection loops increase total generation time; a chapter with 3 rejections takes 3x longer than one approved on first pass","No built-in escalation or timeout mechanism; rejected outputs can loop indefinitely if feedback is contradictory","Requires humans to be available during generation; asynchronous workflows (humans review next day) are not well-supported"],"requires":["Node.js 18+","Human reviewer(s) with access to the CLI, TUI, or Studio Web interface","Clear approval criteria and feedback templates defined per project"],"input_types":["text (agent-generated output awaiting approval)","text (human feedback and revision requests)"],"output_types":["JSON (approval/rejection decision with timestamp)","text (approved output locked for downstream use)","JSON (feedback log for agent learning)"],"categories":["automation-workflow","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_5","uri":"capability://memory.knowledge.chapter.level.snapshot.and.rewriting.with.branching","name":"chapter-level snapshot and rewriting with branching","description":"Creates immutable snapshots of each completed chapter that capture the full text, metadata (word count, character appearances, plot events), and state deltas (what changed in Truth Files). Allows writers to rewind to any previous chapter snapshot and rewrite from that point forward, creating alternate branches without losing the original timeline. The system tracks which snapshots are 'canonical' (approved for publication) vs experimental branches, enabling non-destructive iteration on story direction.","intents":["I want to try a different plot direction starting from chapter 50 without losing the original chapters 1-49","I need to rewrite chapter 30 because I realized a continuity error, and have all downstream chapters automatically re-audited","I want to maintain multiple story branches (e.g., different endings) and compare them side-by-side"],"best_for":["writers experimenting with multiple story directions or endings","projects with long revision cycles where chapters are rewritten multiple times","collaborative projects where different team members work on alternate branches"],"limitations":["Snapshot storage grows linearly with chapter count and branch count; a 500-chapter novel with 5 branches requires ~50-100MB of storage","Re-auditing downstream chapters after a rewrite is computationally expensive; rewriting chapter 50 in a 200-chapter novel requires re-auditing 150 chapters","Branching creates ambiguity about which timeline is 'canonical'; no automatic merge conflict resolution if branches diverge significantly","Snapshots are immutable; correcting a snapshot requires creating a new branch, leading to snapshot proliferation"],"requires":["Node.js 18+","File system storage for snapshots (local or cloud)","Sufficient disk space (1-2MB per chapter per branch)"],"input_types":["text (chapter content to snapshot)","JSON (metadata: character appearances, plot events)"],"output_types":["JSON (snapshot manifest with metadata and state deltas)","text (chapter content from any snapshot)","JSON (branch tree showing relationships between snapshots)"],"categories":["memory-knowledge","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_6","uri":"capability://safety.moderation.genre.specific.writing.rules.and.style.enforcement","name":"genre-specific writing rules and style enforcement","description":"Maintains genre-specific configuration profiles (e.g., 'Xianxia', 'Romance', 'Hard Sci-Fi') that define universal writing rules (prose style, dialogue conventions, pacing expectations) and genre-specific constraints (e.g., Xianxia requires cultivation levels, power systems, sect hierarchies). Rules are stored as markdown documents that agents read before generation, and the Auditor agent validates generated text against these rules, flagging violations. Writers can define custom rules per project and override genre defaults.","intents":["I want to enforce consistent pacing and prose style across all chapters without manually reviewing every one","I need agents to understand genre conventions (e.g., Xianxia cultivation levels) so they don't break world-building rules","I want to define custom writing rules for my specific story (e.g., 'no time travel', 'all dialogue must be in English') and have agents follow them"],"best_for":["genre fiction writers (Xianxia, LitRPG, Romance) with established conventions","publishing teams managing multiple books in the same universe with consistent style requirements","writers who want to enforce personal style preferences (e.g., 'no adverbs', 'short sentences') across all generated content"],"limitations":["Rule enforcement is heuristic-based; complex rules (e.g., 'dialogue must feel natural') are difficult to encode and validate automatically","Overly strict rules can constrain agent creativity; agents may produce repetitive or formulaic text if rules are too prescriptive","Rules are stored as markdown; no formal grammar or schema, making them ambiguous and prone to misinterpretation by agents","No automatic rule conflict detection; contradictory rules (e.g., 'use short sentences' and 'write descriptive prose') can confuse agents"],"requires":["Node.js 18+","Genre profile YAML/JSON files or markdown documents defining rules","Clear, unambiguous rule definitions (ambiguous rules produce poor results)"],"input_types":["markdown (writing rules, genre conventions)","YAML/JSON (genre profile configuration)","text (generated chapters to validate against rules)"],"output_types":["JSON (rule violation report with flagged passages)","text (revised chapters with rule violations corrected)"],"categories":["safety-moderation","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_7","uri":"capability://automation.workflow.inkos.studio.web.workbench.for.visual.project.management","name":"inkos studio web workbench for visual project management","description":"A local web application (built with Vite, React, Hono, Tailwind, Zustand) that provides a visual dashboard for managing novels, viewing real-time generation progress, editing Truth Files, and approving/rejecting agent outputs. The Studio runs on localhost and communicates with the core engine via a REST API, allowing non-technical users to interact with the system without using the CLI. Features include drag-and-drop chapter reordering, inline Truth File editing with live validation, and a visual pipeline status monitor showing which agents are currently running.","intents":["I want a visual interface to manage my novels without using the command line","I need to edit Truth Files (character profiles, world details) through a web UI with validation feedback","I want to see real-time progress as agents generate chapters and approve/reject outputs with a single click"],"best_for":["non-technical writers who prefer GUI over CLI","teams where some members use the CLI and others use the web interface","projects requiring visual collaboration (e.g., showing chapter progress to stakeholders)"],"limitations":["Studio is local-only (localhost); no built-in cloud sync or multi-user collaboration","Web UI adds ~200-300ms latency per action due to HTTP round-trips vs direct CLI calls","Real-time progress updates use polling (not WebSockets); updates are delayed by 1-2 seconds","Drag-and-drop chapter reordering is visual only; doesn't automatically re-audit downstream chapters for continuity"],"requires":["Node.js 18+","Modern web browser (Chrome, Firefox, Safari, Edge)","InkOS core engine running locally (packages/core)"],"input_types":["text (Truth File content edited in web UI)","JSON (approval/rejection decisions from web UI)"],"output_types":["HTML/CSS (web interface)","JSON (REST API responses)","text (Truth Files updated via web UI)"],"categories":["automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_8","uri":"capability://tool.use.integration.terminal.user.interface.tui.with.natural.language.commands","name":"terminal user interface (tui) with natural language commands","description":"A React-Ink-based terminal dashboard that provides an interactive CLI alternative to the web Studio, featuring a visual pipeline status monitor, chapter browser, and Truth File editor. Supports both slash-commands (e.g., '/write chapter 5') and natural language queries (e.g., 'write the next chapter where the protagonist discovers the treasure'). The TUI includes autocomplete for commands, inline help, and a session history that persists across restarts.","intents":["I want to interact with InkOS from the terminal without opening a web browser","I need quick access to common commands (write chapter, review output, edit character) with autocomplete","I want to use natural language to instruct agents instead of remembering exact CLI syntax"],"best_for":["developers and power users comfortable with terminal interfaces","writers who want to stay in their text editor and use the TUI for agent interaction","CI/CD pipelines that need to interact with InkOS programmatically"],"limitations":["TUI rendering is slower than web UI; complex dashboards with many chapters can lag on slower terminals","Natural language command parsing is heuristic-based; ambiguous commands may be misinterpreted","No mouse support; navigation is keyboard-only, which can be slower for some users","Terminal width/height constraints limit the amount of information displayed; large Truth Files must be scrolled"],"requires":["Node.js 18+","Terminal emulator with ANSI color support (most modern terminals)","React-Ink 4.0+ (included in packages/cli)"],"input_types":["text (slash-commands, natural language queries)","keyboard input (navigation, editing)"],"output_types":["ANSI text (terminal UI rendering)","text (command output, logs)"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-narcooo--inkos__cap_9","uri":"capability://tool.use.integration.daemon.mode.with.webhook.system.for.external.integrations","name":"daemon mode with webhook system for external integrations","description":"Allows InkOS to run as a background daemon that listens for webhooks from external systems (e.g., a web novel platform, Discord bot, or custom application). When a webhook is received (e.g., 'generate next chapter'), the daemon triggers the appropriate agent pipeline and sends the result back via HTTP callback. Supports custom webhook handlers that can transform external requests into InkOS commands, enabling integration with third-party platforms without modifying core code.","intents":["I want to trigger chapter generation from my web novel platform's admin panel via webhook","I need to integrate InkOS with a Discord bot so users can request chapters via Discord commands","I want to build a custom application that uses InkOS as a backend service"],"best_for":["web novel platforms (Webnovel, Wattpad, Royal Road) that want to offer AI-assisted writing","Discord communities or Telegram bots that want to generate stories on-demand","developers building custom applications that need novel generation as a service"],"limitations":["Webhook processing adds ~500ms-1s latency per request due to HTTP overhead and request queuing","No built-in authentication or rate limiting; webhooks can be abused to spam generation requests","Webhook callbacks are fire-and-forget; no built-in retry logic if the callback URL is unreachable","Daemon mode requires the InkOS process to run continuously; no built-in process management or auto-restart"],"requires":["Node.js 18+","Network access (daemon must be reachable from external systems)","Webhook URL from external system (e.g., Discord bot, web platform)"],"input_types":["JSON (webhook payload from external system)","text (custom webhook handler code)"],"output_types":["JSON (HTTP response to webhook)","text (generated chapter sent via callback)"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":47,"verified":false,"data_access_risk":"high","permissions":["Node.js 18+","TypeScript 5.0+","API keys for at least one LLM provider (OpenAI, Anthropic, Ollama, or compatible)","Minimum 2GB RAM for local state management and agent coordination","File system write access (local or mounted network drive)","Zod 3.20+ and TypeBox 0.28+ for schema validation","Source material as plain text or markdown","Clear understanding of source canon and desired spinoff direction","High-level story premise (1-2 paragraphs)","Genre profile and writing rules defined"],"failure_modes":["Pipeline latency scales with chapter length and number of agents — typical full cycle (plan→write→audit→revise) takes 2-5 minutes per chapter","Human review gates create blocking dependencies; if a human rejects output, the entire downstream pipeline must restart","Agent coordination overhead adds ~15-20% computational cost vs single-model generation due to inter-agent communication and state synchronization","Requires careful tuning of agent prompts and Truth Files schema per genre; generic configuration produces mediocre results","Truth Files are stored as local markdown/JSON files; no built-in cloud sync or collaborative locking — multiple agents writing simultaneously can cause race conditions","Schema changes require manual migration of existing Truth Files; no automatic schema evolution","File I/O latency adds ~50-100ms per agent read/write cycle; high-frequency state updates (e.g., per-sentence auditing) become bottlenecked","Truth Files grow linearly with chapter count; a 500-chapter novel produces ~2-5MB of state files, requiring careful indexing for fast lookups","Automatic extraction of characters and plot points from source material is imperfect; manual review and correction required","Agents may contradict source canon despite instructions; human review is essential","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.6200159385911297,"quality":0.35,"ecosystem":0.7000000000000001,"match_graph":0.25,"freshness":0.75,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.1,"match_graph":0.28,"freshness":0.12}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-05-24T12:16:22.062Z","last_scraped_at":"2026-05-03T13:57:04.027Z","last_commit":"2026-05-02T10:08:55Z"},"community":{"stars":5520,"forks":1037,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=narcooo--inkos","compare_url":"https://unfragile.ai/compare?artifact=narcooo--inkos"}},"signature":"9gx21DAKOP8LfUmGfzr5By1KnOq0VcfX3REE7liS6KSmfYaYK9TaQR3w4B/z69H9Ql1yEID07J8g1vRWymMmAA==","signedAt":"2026-06-21T09:17:32.574Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/narcooo--inkos","artifact":"https://unfragile.ai/narcooo--inkos","verify":"https://unfragile.ai/api/v1/verify?slug=narcooo--inkos","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}