Skip to content
Back to skills

Repo Recon

ASecurity

Map an unfamiliar repository, service, or module before non-trivial work. Use when entering a codebase not explored in the current session, onboarding, tracing an unknown flow, or recovering from a wrong structural assumption. Do not use for a scoped, obvious change when the relevant files, commands, and conventions are already known.

  • 2 stars
  • 0 votes
  • 0 copies
  • 3 views
  • Added September 19, 2026
ai-agentsrustgit

Security analysis

A100/100

Scanned September 19, 2026

npx -y skills add adityaarakeri/senior-agent-skills --skill repo-recon --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Repo Recon?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Repo Recon
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/adityaarakeri-repo-recon/badge)](https://www.skillsdirectory.com/skills/adityaarakeri-repo-recon)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: repo-recon
description: Map an unfamiliar repository, service, or module before non-trivial work. Use when entering a codebase not explored in the current session, onboarding, tracing an unknown flow, or recovering from a wrong structural assumption. Do not use for a scoped, obvious change when the relevant files, commands, and conventions are already known.
---

# Repo Recon

Map before changing. Keep recon proportional to the task and read-only unless the user explicitly requests a persistent artifact.

## Establish the Boundary

1. Read repository instructions and resolve their scope and precedence.
2. Inspect `git status` or the equivalent before interpreting files. Preserve unrelated and pre-existing work.
3. Keep notes in the agent's ephemeral context. Create `NOTES.md` or another repository file only when the user requests it.

## Map the Relevant System

1. **Read the map.** Inspect the root layout, manifests, lockfiles, build and CI configuration, entry points, generated-code boundaries, and domain docs.
2. **Learn commands safely.** Find install, build, test, lint, and typecheck commands. Inspect cost, side effects, required services, credentials, and likely runtime before executing them. Run the smallest useful baseline and record checks that are unavailable.
3. **Trace one representative flow.** Follow a task-adjacent input through validation, state changes, outputs, error handling, trust boundaries, and observability. Include workers, queues, caches, or persistence when relevant.
4. **Extract conventions.** Note naming, module boundaries, shared helpers, test placement, error patterns, dependency rules, and newer versus legacy patterns.
5. **Report the map.** Summarize relevant paths, exact commands, current baseline, important dependencies, risks, and open questions.

## Standing Rules

- Search before assuming that a file, symbol, command, or configuration exists.
- Read a file before editing it and prefer an established sibling implementation.
- Do not edit generated output when its source or generator is the real boundary.
- Scale the pass to uncertainty and risk, not repository size. Stop when enough evidence exists to perform the requested task safely.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.