agentsope/SkillAlchemy / skills/agentsop-repo-map
agentsop-repo-map
VerifiedSymbol-level code context for LLM coder-agents: tree-sitter extracts symbols, PageRank ranks them over the cross-file reference graph, and the top class/function signatures are fed to the LLM as a token-budgeted read-only map (not RAG, no vector index, human-auditable). Use when an agent must locate the right files in a large/multi-file repo, when builds/refreshes/scopes a repo-map, or when "model edits the wrong file" needs fixing.
| File | Kind | Size |
|---|---|---|
| README.md | markdown | 2.0 KB |
| SKILL.md | skill md | 24.0 KB |
| intermediate/operation_candidates.json | config | 4.9 KB |
| references/R1-source-evidence.md | markdown | 5.3 KB |
| references/R2-cross-tool-comparison.md | markdown | 6.3 KB |
README.md
View on GitHub# repo-map skill
Phase D / C1 of the 7-SOP coder-agent skill landscape.
This skill distills **symbol-level repo-map context** as a reusable, framework-agnostic primitive for LLM-driven code editing. The canonical implementation is Aider's tree-sitter + PageRank repo-map, which hit **70.3% correct-file-identification on SWE-Bench Lite** with no embeddings, no code execution, and no network calls.
## Why this is a standalone skill
Most coder-agent harnesses re-invent some form of "give the LLM a map of the codebase". Aider's experiment is the most thoroughly benchmarked public design. Extracting it as a reusable skill lets:
- Custom agent authors lift the mental model without lifting Aider's CLI.
- Cursor / Cline / Continue users understand what they're (or aren't) getting from their built-in indexers.
- Reviewers audit what the LLM actually saw at decision time.
## Files
- `SKILL.md` — 7-section operational skill. Front-matter `name: repo-map`, `version: 0.1.0`.
- `references/R1-source-evidence.md` — verbatim source quotes with citations.
- `references/R2-cross-tool-comparison.md` — deeper comparison vs RAG, Cursor, Cline.
- `intermediate/operation_candidates.json` — operation-extraction working notes.
## How to use
Read `SKILL.md` end-to-end before the first multi-file edit task. For day-to-day reference jump to:
- §3 SOP — bootstrap / use / refresh / scope cycles.
- §4 operations — 8 named ops with Trigger/Action/Output/Evidence.
- §5 dilemmas — repo-too-big and wrong-file-edited decision trees.
## Citation policy
All non-obvious factual claims cite an Aider source URL inline (`[aider.chat/...]`). The 70.3% figure traces to `aider.chat/2024/05/22/swe-bench-lite.html`. The 25k-token attention degradation traces to `aider.chat/docs/troubleshooting/edit-errors.html`.
## Status
`v0.1.0` — first published distillation. Not yet validated against non-Aider harnesses; the design choices generalize but operation names are conventions, not standards.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |