agentleFS
AurealibeAurealibe/claude-config/.claude/skills/explore/SKILL.md

explore

Skill68 starsChanged 4 months agoMITView on GitHub

Deep codebase exploration to answer specific questions

69 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use agent-context-isolation, agentica-infrastructure and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 8 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 60 edited versions we read keep these.

  • 8

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    After exploration results: reflect on completeness, identify gaps

    line 15
  2. Distinctive:

    Before answering: consider multiple interpretations, ensure accuracy

    line 16
  3. Distinctive:

    Use AskUserQuestion if the question could have multiple interpretations:

    line 32
  4. Distinctive:

    If question is clear and specific, skip to section 2.

    line 37
  5. Distinctive:

    No ceiling, real floor: Use at least 2 agents per domain the question touches (1 only for a genuinely trivial single-file lookup).

    line 48
  6. Distinctive:

    After agents return, verify coverage:

    line 73
  7. Very specific:

    Similar patterns identified?: Do I have a reference implementation?

    line 76
  8. Very specific:

    Data model complete?: If not -> launch additional parallel explore-db agents

    line 77
  9. Very specific:

    Library docs sufficient?: If not -> launch additional parallel explore-docs agents

    line 78
  10. Distinctive:

    CITE: - always reference file:line

    line 115
  11. Distinctive:

    READ-ONLY: - do not modify anything

    line 116

“Common” is measured: kept by at least 6 in 10 of the 60 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

YOU ARE EXECUTING THE /explore SKILL. The user triggered this skill. Follow ALL instructions below step by step. Do NOT treat this as a freeform conversation - execute the skill workflow.

Read-only exploration task.

Ultra Think Strategy

Ultra think before answering:

  • After exploration results: reflect on completeness, identify gaps
  • Before answering: consider multiple interpretations, ensure accuracy
  • Cross-reference findings from different sources

1. PARSE QUESTION

Extract from $ARGUMENTS:

  • Key terms and concepts to search
  • Whether backend, frontend, or both
  • Whether external documentation is needed

1.5 CLARIFY IF AMBIGUOUS

Use AskUserQuestion if the question could have multiple interpretations:

  • "How does X work?" -> Which layer? What aspect?
  • "Where is X handled?" -> At which level?
  • "Find X" -> Find what exactly? (usage/definition/all)

If question is clear and specific, skip to section 2.


2. EXPLORE (PARALLEL)

Launch focused agents in a single message (parallel execution). There is no upper limit on agent count: one agent per distinct angle, scaling up with the question's breadth. The lists below are a starting menu, not a quota.

Agent-count principle

  • You decide the count from the actual scope. Assess how many distinct concerns / areas / file-clusters the question really spans, and launch one focused agent for each. Do not anchor to a fixed number or to file counts.
  • No ceiling, real floor. Use at least 2 agents per domain the question touches (1 only for a genuinely trivial single-file lookup). Above that floor there is no maximum.
  • One distinct focus per agent. Split by angle so two agents never run the same search.
  • No batching. Launch them all in one message.

Backend (one per distinct angle, no maximum):

  1. explore-codebase: "Find [keywords] definitions and core logic in backend/internal/domain/ and backend/internal/application/"
  2. explore-codebase: "Find [keywords] implementations and usage in backend/internal/infrastructure/ and backend/internal/presentation/"
  3. explore-codebase: "Find similar patterns to [keywords] in backend/ for comparison"

Frontend (one per distinct angle, no maximum):

  1. explore-codebase: "Find [keywords] components and hooks in frontend/src/components/ and frontend/src/hooks/"
  2. explore-codebase: "Find [keywords] types, API calls, and pages in frontend/src/types/, frontend/src/lib/, and frontend/src/app/"
  3. explore-codebase: "Find similar patterns to [keywords] in frontend/src/ for comparison"

Supporting (each fans out the same way, one per distinct concern, no maximum):

  • Database: explore-db - "[env] - [tables/schema question]" (one per table cluster / env)
  • Library: explore-docs - "[library] [feature]" (one per library / feature)
  • External: websearch - "[topic]" (one per topic)

For simple single-file lookups, 1 agent per domain is sufficient.


2.5 POST-EXPLORATION CHECK

After agents return, verify coverage:

  1. Full code path traced? Can I trace the complete flow? If gaps -> launch targeted explore-codebase
  2. Similar patterns identified? Do I have a reference implementation? If not -> launch explore-codebase
  3. Data model complete? If not -> launch additional parallel explore-db agents
  4. Library docs sufficient? If not -> launch additional parallel explore-docs agents

3. ANSWER

Provide comprehensive response:

## Answer

[Direct answer to the question]

## Evidence

### Code Found
- `path/to/file.ts:XX` - [description]
- `path/to/file.go:YY` - [description]

### Patterns Identified
- [Pattern 1]: [explanation]

### Documentation (if explore-docs used)
- [Library]: [relevant info]

### External Sources (if websearch used)
- [Title](URL) - [what we learned]

## Recommendations (if applicable)
- [Suggestion based on findings]

Rules

  • PARALLEL - launch relevant agents in ONE message
  • CITE - always reference file:line
  • READ-ONLY - do not modify anything
  • THOROUGH - gather complete context before answering

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.