{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"vscode-merico-devchat","slug":"devchat","name":"DevChat","type":"extension","url":"https://marketplace.visualstudio.com/items?itemName=merico.devchat","page_url":"https://unfragile.ai/devchat","categories":["code-editors"],"tags":["ai","anthropic","assistant","autocomplete","bot","c#","c++","chat","chatbot","codegen","codex","co-pilot","deepseek","deepseek-r1","deepseek-v3","devchat","documentation","go","golang","intellisense","java","javascript","keybindings","kotlin","llm","model","openai","php","pilot","python","refactor","ruby","snippets","sourcegraph","test","typescript"],"pricing":{"model":"freemium","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"vscode-merico-devchat__cap_0","uri":"capability://code.generation.editing.prompt.centric.code.generation.with.manual.context.selection","name":"prompt-centric code generation with manual context selection","description":"DevChat generates code by accepting natural language prompts paired with explicitly selected code context. Unlike auto-completion tools that infer context automatically, DevChat requires developers to manually select relevant code snippets, file contents, git diffs, and command outputs to include in the prompt before sending to the LLM. This manual context assembly workflow is stored as reusable prompt templates in the ~/.chat/workflows/ directory structure (sys/, org/, usr/ subdirectories), enabling reproducible code generation patterns without requiring complex prompt engineering frameworks.","intents":["Generate new code functions from natural language descriptions while controlling exactly what context the AI sees","Create code variants by reusing prompt templates with different selected code snippets","Build reproducible code generation workflows that can be shared across teams via the org/ prompt directory"],"best_for":["developers who prefer explicit control over AI context over automatic inference","teams building standardized code generation workflows with reusable prompts","solo developers prototyping features where context precision matters more than speed"],"limitations":["Requires manual context selection for each prompt — no automatic context inference means slower workflow than inline completion tools","No real-time inline code completion — designed for discrete prompt-response cycles rather than continuous suggestions","Prompt template system is text-file based with no built-in validation or schema enforcement — malformed templates may fail silently"],"requires":["Visual Studio Code (minimum version unknown — not documented)","LLM API access (provider and configuration mechanism unknown — not documented)","Write permissions to ~/.chat/workflows/ directory for custom prompt storage"],"input_types":["natural language prompts","selected code snippets from editor","file paths and directory structures","git diff output (via git diff --cached command integration)","shell command outputs (arbitrary command results can be piped into prompts)"],"output_types":["generated code (language-agnostic based on prompt context)","code snippets ready for insertion into editor"],"categories":["code-generation-editing","prompt-engineering"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_1","uri":"capability://code.generation.editing.context.aware.test.case.generation.from.existing.patterns","name":"context-aware test case generation from existing patterns","description":"DevChat analyzes existing test cases in the project and generates new test cases for functions by referencing the discovered test patterns and conventions. The extension extracts test file structure, assertion patterns, and testing framework usage from the codebase, then incorporates this context into prompts to generate tests that match the project's established testing style. This pattern-matching approach ensures generated tests follow local conventions rather than imposing a generic testing style.","intents":["Generate test cases for new functions that match the project's existing testing patterns and framework","Understand what testing conventions are used in the codebase and apply them to new code","Create comprehensive test suites by learning from existing test examples in the project"],"best_for":["teams with established testing conventions who want generated tests to match local style","projects using non-standard testing frameworks where generic test generation fails","developers who want to maintain test consistency across a codebase"],"limitations":["Requires existing test cases in the project to learn patterns from — projects with no tests cannot use this capability","Pattern extraction mechanism is undocumented — unclear how deeply it analyzes test structure vs. simple regex matching","No validation that generated tests actually pass or cover the target function — relies on developer review"],"requires":["Existing test files in the project (format and framework unknown — not documented)","Target function or code to generate tests for","Manual selection of reference test cases to include in the prompt context"],"input_types":["function or code block to test","existing test file examples (selected manually)","test framework and assertion patterns (extracted from codebase)"],"output_types":["generated test code matching project conventions","test cases ready for insertion into test files"],"categories":["code-generation-editing","testing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_2","uri":"capability://code.generation.editing.git.aware.commit.message.generation.from.staged.changes","name":"git-aware commit message generation from staged changes","description":"DevChat integrates with git to analyze staged changes (via git diff --cached) and generates commit messages that describe the modifications. The extension reads the diff output, analyzes the code changes, and produces commit messages that summarize what was changed and why. This capability bridges the gap between code changes and human-readable commit history by using the actual diff as context for message generation.","intents":["Generate descriptive commit messages from staged code changes without manually writing them","Ensure commit messages accurately reflect the actual code modifications made","Maintain consistent commit message style across a team by using standardized prompts"],"best_for":["developers working in git repositories who want to automate commit message writing","teams with commit message conventions who want to enforce style via prompt templates","projects where commit history quality matters (open source, regulated industries)"],"limitations":["Requires git repository with staged changes — cannot generate messages for unstaged or committed changes","Relies on diff analysis which may miss context about why changes were made (motivation, design decisions)","No integration with git hooks — requires manual invocation rather than automatic pre-commit generation"],"requires":["Git repository initialized in the workspace","Changes staged with git add (git diff --cached must return non-empty output)","Manual invocation of the commit message generation prompt"],"input_types":["git diff output from staged changes","optional: custom commit message template/prompt"],"output_types":["generated commit message text","message ready for git commit -m"],"categories":["code-generation-editing","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_3","uri":"capability://text.generation.language.code.explanation.with.dependency.aware.context.extraction","name":"code explanation with dependency-aware context extraction","description":"DevChat explains code by analyzing the selected code block and automatically extracting definitions of dependent functions and symbols that are referenced. When a developer selects a function to explain, the extension identifies external function calls, class references, and imported symbols, then includes their definitions in the prompt context sent to the LLM. This dependency-aware approach ensures explanations include necessary context without requiring developers to manually hunt down related code.","intents":["Understand how a function works by getting explanations that include definitions of functions it calls","Learn about unfamiliar code patterns by seeing the full dependency chain explained","Onboard to a new codebase faster by understanding code with all its dependencies resolved"],"best_for":["developers new to a codebase who need to understand code with full context","teams with complex interdependencies where manual context gathering is time-consuming","code review workflows where reviewers need to understand unfamiliar code quickly"],"limitations":["Symbol extraction mechanism is undocumented — unclear how it handles dynamic imports, reflection, or indirect dependencies","May include excessive context if a function has many dependencies — no filtering or prioritization of included definitions","Language-specific symbol resolution unknown — may fail for languages with complex scoping rules (e.g., C++ templates, Python metaclasses)"],"requires":["Code selection in the editor (function, class, or code block)","Language server or parser support for the selected code's language (mechanism unknown)","Definitions of dependent symbols must be in the same workspace or accessible files"],"input_types":["selected code block (function, class, or snippet)","language context (inferred from file extension or editor mode)"],"output_types":["natural language explanation of the code","explanation including definitions of dependencies"],"categories":["text-generation-language","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_4","uri":"capability://text.generation.language.documentation.generation.from.code.with.context.control","name":"documentation generation from code with context control","description":"DevChat generates documentation by accepting selected code and optional context (function signatures, type definitions, usage examples) and producing formatted documentation. The extension supports generating documentation in various formats (docstrings, markdown, API docs) based on the prompt template used. Unlike automatic documentation tools, DevChat requires explicit selection of what code to document and what context to include, giving developers control over documentation scope and style.","intents":["Generate docstrings or documentation comments for functions without manually writing them","Create API documentation from code by selecting relevant functions and including usage examples","Maintain consistent documentation style across a project by using standardized documentation prompts"],"best_for":["teams with documentation standards who want to enforce style via prompt templates","open source projects where documentation quality affects adoption","developers who want to document code but lack time for manual writing"],"limitations":["Requires manual selection of code to document — no automatic discovery of undocumented functions","Documentation accuracy depends on code clarity — poorly written code produces poor documentation","No validation that generated documentation is accurate or complete — relies on developer review"],"requires":["Code selection in the editor (function, class, or module)","optional: usage examples or context to include in the prompt","Documentation template/prompt (can use default or custom)"],"input_types":["selected code (function, class, module, or file)","optional: usage examples, type information, or additional context","documentation format preference (via prompt template)"],"output_types":["generated documentation (docstring, markdown, or custom format)","documentation ready for insertion into code or docs files"],"categories":["text-generation-language","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_5","uri":"capability://memory.knowledge.extensible.filesystem.based.prompt.workflow.system","name":"extensible filesystem-based prompt workflow system","description":"DevChat stores and manages prompts as text files in a hierarchical directory structure (~/.chat/workflows/) organized into sys/ (system prompts), org/ (organization-level), and usr/ (user-level) directories. Prompts are plain text files that can be edited with any text editor, version-controlled in git, and shared across teams. This filesystem-based approach treats prompts as code artifacts rather than ephemeral chat history, enabling teams to build prompt libraries and standardize AI interactions without proprietary prompt management tools.","intents":["Create reusable prompts that can be shared across a team and version-controlled in git","Build a library of standardized prompts for common code generation tasks","Customize prompts for organization-specific coding standards and conventions"],"best_for":["teams building standardized AI-assisted workflows","organizations wanting to version-control and audit AI prompts","developers who prefer text-based configuration over UI-driven settings"],"limitations":["No built-in prompt validation or schema enforcement — malformed prompts may fail silently","No UI for prompt management — requires manual file editing and directory organization","No conflict resolution for overlapping prompts in sys/org/usr hierarchy — precedence rules unknown","Prompt template syntax is undocumented — unclear what variables or interpolation is supported"],"requires":["Write permissions to ~/.chat/workflows/ directory","Text editor for creating/editing prompt files","Understanding of the directory hierarchy (sys/org/usr) and naming conventions (unknown)"],"input_types":["plain text prompt templates","optional: variable placeholders or interpolation syntax (undocumented)"],"output_types":["organized prompt files in ~/.chat/workflows/","prompts ready for use in code generation workflows"],"categories":["memory-knowledge","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_6","uri":"capability://tool.use.integration.command.output.integration.into.prompt.context","name":"command output integration into prompt context","description":"DevChat allows developers to include arbitrary shell command outputs in prompts by executing commands (e.g., git diff --cached, tree ./src, npm list) and capturing their output as context. This capability enables prompts to reference dynamic information about the project state (file structure, dependencies, git status) without requiring manual copy-paste. The extension executes commands in the workspace context and includes the output in the prompt sent to the LLM.","intents":["Include project structure information in prompts by running tree or find commands","Reference git status and changes by running git commands and including their output","Incorporate dependency information by running package manager commands (npm list, pip freeze, etc.)"],"best_for":["developers who want to reference dynamic project state in AI prompts","teams using complex project structures where manual context gathering is error-prone","workflows where code generation depends on current git state or dependency versions"],"limitations":["Command execution happens in the workspace context — may fail if commands are not available or workspace is misconfigured","No sandboxing or safety restrictions documented — unclear if arbitrary commands can be executed","Command output size is unbounded — large outputs (e.g., tree of a huge project) may exceed LLM context limits","No caching of command outputs — each prompt execution re-runs the commands, adding latency"],"requires":["Shell access in the workspace (bash, sh, or Windows equivalent)","Commands to be available in the workspace environment (git, tree, npm, etc.)","Workspace folder open in VS Code"],"input_types":["shell commands as text in prompt templates","workspace context (current directory, environment variables)"],"output_types":["command output captured as text","output included in prompt context sent to LLM"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_7","uri":"capability://code.generation.editing.multi.language.code.generation.with.language.agnostic.prompts","name":"multi-language code generation with language-agnostic prompts","description":"DevChat generates code for multiple programming languages (Python, JavaScript, TypeScript, Java, C++, C#, Go, Kotlin, PHP, Ruby) using the same prompt interface. The extension infers the target language from the editor context (file extension, language mode) and includes language-specific context (syntax, conventions, frameworks) in the prompt. This language-agnostic prompt interface allows developers to write prompts once and apply them across different languages without language-specific prompt variants.","intents":["Generate code in different languages using the same prompt by changing the file context","Create polyglot projects where code generation works consistently across languages","Learn new languages by generating code and seeing language-specific patterns"],"best_for":["polyglot teams working across multiple languages","developers learning new languages who want AI assistance","projects with code generation pipelines that target multiple languages"],"limitations":["Language detection is automatic but may fail for ambiguous file types (e.g., .js could be Node.js or browser JavaScript)","Language-specific context inclusion is undocumented — unclear how deeply it understands language conventions","Code generation quality varies by language — LLM may be better trained on some languages than others","No language-specific validation or linting — generated code may not follow language best practices"],"requires":["File open in editor with recognized language extension","LLM support for the target language (varies by model)"],"input_types":["natural language prompts (language-agnostic)","code context in any supported language"],"output_types":["generated code in the target language","code snippets matching the file's language"],"categories":["code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_8","uri":"capability://tool.use.integration.vs.code.editor.integration.with.manual.context.selection.ui","name":"vs code editor integration with manual context selection ui","description":"DevChat integrates with VS Code through an extension that provides UI for selecting code context and invoking prompts. The extension likely uses VS Code's editor selection API to capture highlighted code, command palette for prompt invocation, and sidebar or panel UI for displaying results. The manual context selection workflow requires developers to explicitly highlight code, select reference files, and invoke prompts rather than relying on automatic context inference. Integration points are undocumented but likely include editor selection, command palette, and possibly context menus.","intents":["Select code in the editor and send it to AI prompts without leaving VS Code","Invoke code generation workflows using keyboard shortcuts or command palette","View AI-generated code results in the editor and insert them into files"],"best_for":["VS Code users who want AI assistance without switching applications","developers who prefer keyboard-driven workflows (command palette)","teams standardizing on VS Code as the development environment"],"limitations":["VS Code-only — no support for other editors (JetBrains IDEs, Vim, Emacs, etc.)","UI/UX details are undocumented — unclear how results are displayed or inserted","Keybindings and command palette commands are undocumented — users must discover them","No multi-file selection UI documented — unclear how to select context across multiple files"],"requires":["Visual Studio Code (minimum version unknown)","DevChat extension installed from VS Code Marketplace","VS Code workspace open with files to work with"],"input_types":["editor selection (highlighted code)","command palette invocation","optional: context menu selection"],"output_types":["generated code displayed in editor or panel","code ready for insertion into files"],"categories":["tool-use-integration","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"vscode-merico-devchat__cap_9","uri":"capability://code.generation.editing.refactoring.code.generation.with.context.aware.suggestions","name":"refactoring code generation with context-aware suggestions","description":"DevChat can generate refactored code by accepting selected code and a refactoring intent (simplify, optimize, improve readability, etc.) and producing refactored versions. The extension includes the original code and optional context (related functions, test cases, performance constraints) in the prompt to guide the refactoring. Unlike automated refactoring tools, DevChat's approach is prompt-based and requires explicit intent specification, giving developers control over refactoring style and scope.","intents":["Simplify complex code by generating cleaner, more readable versions","Optimize code performance by generating faster implementations","Improve code maintainability by refactoring for clarity and consistency"],"best_for":["developers who want AI-assisted refactoring with explicit control over changes","teams with refactoring standards who want to enforce style via prompts","code review workflows where reviewers suggest refactoring via AI"],"limitations":["No automated refactoring — requires manual prompt invocation for each refactoring","No validation that refactored code is functionally equivalent — relies on developer testing","Refactoring quality depends on LLM understanding of code semantics — may introduce subtle bugs","No integration with test suites — cannot automatically verify that refactoring preserves behavior"],"requires":["Code selection in the editor (function, class, or code block)","Refactoring intent specified in the prompt","optional: test cases or performance constraints to guide refactoring"],"input_types":["selected code to refactor","refactoring intent (simplify, optimize, improve readability, etc.)","optional: test cases, performance constraints, or style guidelines"],"output_types":["refactored code","code ready for testing and review"],"categories":["code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":38,"verified":false,"data_access_risk":"high","permissions":["Visual Studio Code (minimum version unknown — not documented)","LLM API access (provider and configuration mechanism unknown — not documented)","Write permissions to ~/.chat/workflows/ directory for custom prompt storage","Existing test files in the project (format and framework unknown — not documented)","Target function or code to generate tests for","Manual selection of reference test cases to include in the prompt context","Git repository initialized in the workspace","Changes staged with git add (git diff --cached must return non-empty output)","Manual invocation of the commit message generation prompt","Code selection in the editor (function, class, or code block)"],"failure_modes":["Requires manual context selection for each prompt — no automatic context inference means slower workflow than inline completion tools","No real-time inline code completion — designed for discrete prompt-response cycles rather than continuous suggestions","Prompt template system is text-file based with no built-in validation or schema enforcement — malformed templates may fail silently","Requires existing test cases in the project to learn patterns from — projects with no tests cannot use this capability","Pattern extraction mechanism is undocumented — unclear how deeply it analyzes test structure vs. simple regex matching","No validation that generated tests actually pass or cover the target function — relies on developer review","Requires git repository with staged changes — cannot generate messages for unstaged or committed changes","Relies on diff analysis which may miss context about why changes were made (motivation, design decisions)","No integration with git hooks — requires manual invocation rather than automatic pre-commit generation","Symbol extraction mechanism is undocumented — unclear how it handles dynamic imports, reflection, or indirect dependencies","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.42,"quality":0.3,"ecosystem":0.35000000000000003,"match_graph":0.25,"freshness":0.75,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.15,"match_graph":0.23,"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=devchat","compare_url":"https://unfragile.ai/compare?artifact=devchat"}},"signature":"JMgCzeBVCUTmOFer3c8GNbMJyF+HiSvNFgQ86Ic58n+pMgpviSlsKrs7O3BvnSfTMjdkxw5LYWFM4bVGBtk/Ag==","signedAt":"2026-06-19T22:52:11.599Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/devchat","artifact":"https://unfragile.ai/devchat","verify":"https://unfragile.ai/api/v1/verify?slug=devchat","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"}}