Agentlens

**Quick judgement:** Agentlens is a focused, documentation-driven navigation skill for large or complex codebases that already have an `.agentlens` directory. It is well-suited for structured exploration and impact-aware editing but depends heavily on the quality and freshness of the pre-generated docs. **Strengths:** - Enforces a clear navigation flow: `INDEX.md → MODULE.md → outline/memory/imports → source`, which can significantly speed up understanding of big projects. - Provides specialized views: overviews, symbol outlines, TODOs/warnings/rules, and dependency maps. - Minimizes the need to open or scan large source files directly, which is especially useful for monoliths or repositories with very large files. **Key risks / limitations:** - **Stale or incomplete docs:** If the `.agentlens` documentation hasn’t been regenerated recently, the structure, symbols, or TODOs may not match the current code, leading to confusion or incorrect assumptions. - **Requires Agentlens setup:** The skill is only truly effective in projects that have been processed by Agentlens and follow the documented layout (`INDEX.md`, `modules/*`, `files/*`). For unprocessed repos, it adds little value. - **Coverage gaps:** Some files or modules may be missing detailed docs, so you may still need to fall back to raw source inspection. **Recommended scenarios:** - Exploring a new, medium-to-large codebase where an `.agentlens` directory is present and reasonably up to date. - Planning changes or refactors while reviewing `memory.md` for warnings, business rules, and TODOs, and `imports.md` for dependency impact. - Quickly finding where a symbol or feature lives in a sprawling repository by following `INDEX.md` → module docs → `outline.md`. - Code review or architecture review sessions where a fast, high-level structural map of the project is needed before diving into individual files.