{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"vscode-sourcegraph-amp","slug":"amp-research-preview","name":"Amp (Research Preview)","type":"agent","url":"https://marketplace.visualstudio.com/items?itemName=sourcegraph.amp","page_url":"https://unfragile.ai/amp-research-preview","categories":["ai-agents"],"tags":["agent","ai","amp","amp code","ampcode","assistant","Claude","Claude Sonnet","Claude Sonnet 4.6","free","GPT-5","jsonc","keybindings","llm","sourcegraph"],"pricing":{"model":"freemium","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"vscode-sourcegraph-amp__cap_0","uri":"capability://code.generation.editing.multi.model.agentic.code.generation.with.mode.based.routing","name":"multi-model agentic code generation with mode-based routing","description":"Generates new code from natural language requests by routing to different LLM backends based on user-selected mode: 'smart' mode uses Claude Opus 4.6 or GPT-5.4 for complex reasoning, 'rush' mode uses Claude Haiku 4.5 for fast execution, and 'deep' mode uses GPT-5.3 Codex with extended thinking for complex problem-solving. The agent maintains conversation threads within VS Code, allowing users to iteratively refine generated code through multi-turn dialogue without losing context.","intents":["Generate boilerplate or new functions from natural language descriptions","Create entire files or modules based on requirements","Switch between speed and quality tradeoffs for different coding tasks","Maintain conversation history across multiple code generation requests"],"best_for":["solo developers building features quickly","teams wanting to standardize on Claude or GPT models without managing separate API keys","developers who want to toggle between fast iteration and deep reasoning without context loss"],"limitations":["Model selection mechanism is undocumented — unclear if modes are user-selectable per-request or globally configured","Model names (GPT-5.4, Claude Opus 4.6) do not match current public model naming conventions, suggesting documentation may be aspirational or outdated","No documented support for custom models or local LLM backends","Conversation threads are stored server-side with unknown retention policy and privacy guarantees","Research Preview status means API and model routing logic may change without notice"],"requires":["VS Code (minimum version unknown)","Active internet connection for API calls to model providers","Authentication via ampcode.com/install web flow","Sourcegraph account (implied by authentication requirement)"],"input_types":["natural language text requests","code context from current file or codebase"],"output_types":["generated code (functions, classes, modules, files)","code explanations in natural language"],"categories":["code-generation-editing","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_1","uri":"capability://code.generation.editing.codebase.aware.multi.file.code.modification.with.human.review.workflow","name":"codebase-aware multi-file code modification with human review workflow","description":"Modifies existing code across multiple files in the user's codebase by analyzing project structure and context, then presenting all proposed changes in a built-in review panel before application. The agent understands the full codebase scope (not just the current file) and can coordinate edits across related files. Changes are held in a staging state until the user explicitly approves them, preventing accidental overwrites.","intents":["Refactor code across multiple files while maintaining consistency","Apply automated fixes or updates to the entire codebase","Review AI-generated changes before committing them to disk","Understand what the agent plans to change before accepting modifications"],"best_for":["teams requiring code review workflows before AI-generated changes are applied","developers working on large codebases where single-file edits are insufficient","projects where accidental overwrites are high-risk (production code, shared files)"],"limitations":["Review panel UI/UX details are undocumented — unclear how diffs are visualized or how granular approval is (per-file, per-change, all-or-nothing)","File system access scope is undefined — unclear if agent can read all files or only those in open workspace","No documented support for version control integration (Git awareness, branch-specific changes, merge conflict handling)","Terminal context integration is mentioned but mechanism is completely undocumented","No documented rollback or undo mechanism if changes are applied and then need reversal"],"requires":["VS Code workspace with one or more files","Active internet connection","Authentication via ampcode.com/install","Sufficient API quota for multi-file analysis and modification"],"input_types":["natural language modification requests","existing code files in the workspace","project structure and file relationships"],"output_types":["modified code files (staged for review)","diff/patch representation of changes","change summary or explanation"],"categories":["code-generation-editing","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_2","uri":"capability://memory.knowledge.persistent.conversation.threading.with.code.context.preservation","name":"persistent conversation threading with code context preservation","description":"Maintains multi-turn conversation threads within the VS Code sidebar, allowing users to iteratively refine code generation and modification requests while preserving full context across turns. Each thread stores the conversation history, generated code, and applied changes, enabling users to reference previous requests and build on prior work without re-explaining context. Threads can be saved and shared (mechanism undocumented).","intents":["Iteratively refine generated code through multiple back-and-forth requests","Build complex features incrementally without losing context between requests","Save conversation history for documentation or knowledge sharing","Share code generation workflows with teammates"],"best_for":["developers building features incrementally over multiple requests","teams documenting how code was generated for knowledge transfer","solo developers who want to preserve their reasoning and iteration process"],"limitations":["Thread storage location and persistence model are undocumented — unclear if stored locally or server-side","Data retention policy is unknown — no documentation on how long threads are kept or whether they can be deleted","Privacy and data handling are undocumented — unclear if threads are encrypted, who can access them, or whether Sourcegraph/model providers can access conversation content","Share mechanism is undocumented — unclear if threads are shared via URL, export, or other method","No documented support for exporting threads to other formats (markdown, JSON, etc.)","Conversation context window limits are unknown — unclear how many turns or tokens can be stored per thread"],"requires":["VS Code with Amp extension installed","Active internet connection","Authentication via ampcode.com/install","Sourcegraph account for thread persistence"],"input_types":["natural language requests","code snippets and context from previous turns","file references and project context"],"output_types":["conversation history (text)","generated code from each turn","thread metadata (creation date, last modified, etc.)"],"categories":["memory-knowledge","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_3","uri":"capability://planning.reasoning.extended.thinking.code.reasoning.for.complex.problem.solving","name":"extended-thinking code reasoning for complex problem-solving","description":"Activates a 'deep' mode that routes requests to GPT-5.3 Codex with extended thinking capabilities, enabling the agent to reason through complex coding problems step-by-step before generating solutions. This mode is designed for problems that require multi-step reasoning, architectural decisions, or deep analysis of existing code. Extended thinking adds latency but produces higher-quality solutions for difficult problems.","intents":["Solve complex architectural or algorithmic problems that require deep reasoning","Analyze and refactor large, complex code sections","Debug subtle bugs that require understanding code flow across multiple files","Design solutions for non-trivial requirements"],"best_for":["developers tackling complex algorithmic or architectural problems","teams debugging subtle issues that require deep code analysis","projects where solution quality is more important than speed"],"limitations":["Extended thinking adds significant latency — no documented response time, but extended thinking models typically add 5-30+ seconds per request","Model name 'GPT-5.3 Codex' does not match current public OpenAI model naming, suggesting documentation is outdated or aspirational","No documented way to control extended thinking depth or reasoning budget","No documented fallback if extended thinking fails or times out","Unclear whether extended thinking is available in free tier or requires paid subscription","No documented examples of when to use 'deep' mode vs. 'smart' mode"],"requires":["VS Code with Amp extension installed","Active internet connection with sufficient bandwidth for extended thinking responses","Authentication via ampcode.com/install","Possible paid subscription (pricing model unknown)"],"input_types":["complex natural language problem descriptions","code snippets or files requiring deep analysis","architectural or design questions"],"output_types":["step-by-step reasoning explanation","generated code solution","architectural recommendations"],"categories":["planning-reasoning","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_4","uri":"capability://tool.use.integration.terminal.integrated.coding.agent.with.undocumented.context.passing","name":"terminal-integrated coding agent with undocumented context passing","description":"Integrates with the VS Code terminal to enable the agent to receive context from terminal output, error messages, and command execution results. The agent can use this terminal context to generate fixes, debug issues, or provide recommendations based on actual runtime behavior. The specific mechanism for passing terminal context to the agent is completely undocumented.","intents":["Generate fixes based on error messages from terminal output","Debug runtime issues by analyzing error traces and logs","Suggest improvements based on test failures or linter output","Automate responses to common terminal errors"],"best_for":["developers who want the agent to see and respond to terminal errors in real-time","teams using CI/CD or test-driven development where terminal output is critical","developers debugging issues where error messages are essential context"],"limitations":["Terminal integration mechanism is completely undocumented — unclear how terminal context is captured, filtered, or passed to the agent","No documented scope limits — unclear if all terminal output is captured or if there are privacy/security filters","No documented latency impact of terminal context passing","Unclear whether terminal integration works with remote terminals (SSH, containers, WSL)","No documented way to disable terminal context passing if privacy is a concern","No examples or documentation of terminal-based workflows"],"requires":["VS Code with integrated terminal","Amp extension installed and authenticated","Active internet connection","Terminal context must be visible in VS Code (not external terminals)"],"input_types":["terminal output (stdout/stderr)","error messages and stack traces","command execution results","test output and logs"],"output_types":["suggested fixes based on errors","code modifications to address issues","explanations of error causes"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_5","uri":"capability://automation.workflow.opinionated.agent.design.with.no.backward.compatibility.guarantee","name":"opinionated agent design with no backward compatibility guarantee","description":"Implements an explicitly opinionated design philosophy that prioritizes forward progress and feature iteration over backward compatibility. The agent makes specific architectural choices about which features to include/exclude and explicitly states 'No backcompat, no legacy features' as a design principle. This allows rapid iteration and feature changes but means breaking changes can occur between versions without deprecation warnings.","intents":["Use a cutting-edge coding agent that prioritizes innovation over stability","Adopt an agent that will evolve quickly with new capabilities","Work with a tool that makes strong architectural choices rather than supporting legacy patterns"],"best_for":["early adopters willing to accept breaking changes for access to frontier features","research teams and developers exploring cutting-edge AI coding capabilities","teams not dependent on stable APIs or long-term backward compatibility"],"limitations":["Breaking changes can occur between versions without deprecation period","Research Preview status means features, APIs, and behavior may change significantly","No documented migration path for breaking changes","Unclear which specific features are considered 'opinionated' vs. configurable","No documented stability guarantees or versioning scheme","May be discontinued without notice (research preview status)"],"requires":["Acceptance of potential breaking changes and instability","Willingness to update extension frequently","No dependency on stable long-term APIs"],"input_types":[],"output_types":[],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_6","uri":"capability://automation.workflow.freemium.model.with.undocumented.paid.tier.and.quota.system","name":"freemium model with undocumented paid tier and quota system","description":"Offers free access to the agent with an undocumented pricing model for advanced features or higher usage. The free tier provides access to the agent's core capabilities, but specific quotas, rate limits, and paid tier features are not documented. The extension is installable at no cost, but usage-based or feature-based pricing may apply.","intents":["Try the agent at no cost before committing to paid usage","Use the agent for personal projects or learning without payment","Understand pricing and quota implications before adopting for team use"],"best_for":["individual developers evaluating the agent before team adoption","open-source projects or non-commercial use","developers with low usage who can stay within free tier limits"],"limitations":["Pricing model is completely undocumented — no information on cost, quotas, or feature tiers","Free tier limits are unknown — unclear if there are request limits, file size limits, or feature restrictions","No documented way to monitor usage or quota consumption","Unclear whether free tier includes all modes (smart/rush/deep) or if deep mode requires paid subscription","No documented upgrade path or pricing comparison","No information on whether free tier is time-limited (trial) or permanent"],"requires":["VS Code with Amp extension installed","Authentication via ampcode.com/install","Possible payment method on file for paid tier (if applicable)"],"input_types":[],"output_types":[],"categories":["automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-sourcegraph-amp__cap_7","uri":"capability://tool.use.integration.vs.code.sidebar.based.agent.ui.with.activity.bar.integration","name":"vs code sidebar-based agent ui with activity bar integration","description":"Provides a dedicated sidebar panel in VS Code for agent interaction, accessible via an Amp icon in the activity bar. The sidebar serves as the primary UI for issuing natural language requests, viewing conversation threads, and managing agent state. This integration keeps the agent accessible without requiring separate windows or applications.","intents":["Access the agent without leaving VS Code or opening additional windows","View conversation history and threads in the sidebar while editing code","Quickly switch between coding and agent interaction"],"best_for":["developers who want agent access integrated into their primary editor","teams using VS Code as their standard development environment","developers who prefer sidebar-based tools over separate chat windows"],"limitations":["Sidebar real estate is limited — unclear how much space is available for conversation threads and code display","No documented keyboard shortcuts or command palette integration for faster access","No documented support for custom keybindings","Unclear whether sidebar can be resized, docked to different positions, or hidden","No documented support for multi-monitor workflows or external displays","Activity bar icon is always visible, potentially cluttering the UI for users who don't use the agent frequently"],"requires":["VS Code (minimum version unknown)","Amp extension installed","Sufficient sidebar space (depends on VS Code window size)"],"input_types":["natural language text input in sidebar","code context from editor"],"output_types":["conversation display in sidebar","code suggestions and modifications"],"categories":["tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":41,"verified":false,"data_access_risk":"high","permissions":["VS Code (minimum version unknown)","Active internet connection for API calls to model providers","Authentication via ampcode.com/install web flow","Sourcegraph account (implied by authentication requirement)","VS Code workspace with one or more files","Active internet connection","Authentication via ampcode.com/install","Sufficient API quota for multi-file analysis and modification","VS Code with Amp extension installed","Sourcegraph account for thread persistence"],"failure_modes":["Model selection mechanism is undocumented — unclear if modes are user-selectable per-request or globally configured","Model names (GPT-5.4, Claude Opus 4.6) do not match current public model naming conventions, suggesting documentation may be aspirational or outdated","No documented support for custom models or local LLM backends","Conversation threads are stored server-side with unknown retention policy and privacy guarantees","Research Preview status means API and model routing logic may change without notice","Review panel UI/UX details are undocumented — unclear how diffs are visualized or how granular approval is (per-file, per-change, all-or-nothing)","File system access scope is undefined — unclear if agent can read all files or only those in open workspace","No documented support for version control integration (Git awareness, branch-specific changes, merge conflict handling)","Terminal context integration is mentioned but mechanism is completely undocumented","No documented rollback or undo mechanism if changes are applied and then need reversal","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.6,"quality":0.26,"ecosystem":0.35000000000000003,"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":"2026-05-24T12:16:34.803Z","last_scraped_at":"2026-05-03T15:20:33.198Z","last_commit":null},"community":{"stars":null,"forks":null,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=amp-research-preview","compare_url":"https://unfragile.ai/compare?artifact=amp-research-preview"}},"signature":"PFWR2XXUlosURE7Ci+cz2IP1s9oVpGDBz1WGCcm1HCCvOhdi7kuUpEIix19TcRP5aHwuBPepoG/hQtzJjtUlCw==","signedAt":"2026-06-23T08:23:17.085Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/amp-research-preview","artifact":"https://unfragile.ai/amp-research-preview","verify":"https://unfragile.ai/api/v1/verify?slug=amp-research-preview","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"}}