repo-context
Skill7 starsChanged 22 days agoMITView on GitHub
Read-only onboarding discovery for an unfamiliar repository or archive. Reports the owning Git root, HEAD and dirty state, instruction files, Node/Python/Go manifests (including mixed and nested layouts), declared toolchains, test declarations and command candidates as data, with malformed or undeclared facts listed as explicit unknowns. Use when starting work in a repository, before choosing build or test commands, or when asked what a project uses. It never executes candidates, guesses a package manager, default branch or CI gate, or reads credentials or the environment.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- aymardkouakou/claude-gh-skillstheir own version, first seen Nov 8, 2026
- hushh-labs/hushh-researchtheir own version, first seen Nov 1, 2026
- justEstif/pi-shelftheir own version, first seen Oct 13, 2026
- Chris-Carruthers/canontheir own version, first seen Aug 29, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- hushh-labs/hushh-research★ 40same name
- Chris-Carruthers/canon★ 4same name
- aymardkouakou/claude-gh-skills★ 1same name
- justEstif/pi-shelf★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository31 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- api-contract-testingSkill
- browser-testingSkill
- ci-cd-templatesSkill
- code-review-checklistSkill
- context-escalationSkill
- deep-analysisSkill
- deep-readSkill
- deliverySkill
- deployment-runbookSkill
- divergeSkill
- docker-deploymentSkill
- executeSkill
- experiment-loopSkill
- git-workflowSkill
- handoffSkill
- infrastructure-as-codeSkill
- investigateSkill
- kubernetes-opsSkill
- library-docsSkill
- multi-agent-orchestrationSkill
- observability-stackSkill
- remote-triggersSkill
- scheduled-tasksSkill
- security-scanSkill
- skill-creatorSkill
- sre-runbooksSkill
- ui-guidelinesSkill
- verified-handoffSkill
- worktree-workflowSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Read every listed instruction file.
- Very specific: line 37
Treat defaultBranch: null, ci.gates: null and packageManager: null as unknowns to confirm from instructions or the user.
- Very specific: line 43
Treat every candidate as unverified until it runs.
“Common” is measured: kept by at least 6 in 10 of the 4 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.
Files
Repo Context
Establish facts about a repository before acting on assumptions about it.
Run
python3 scripts/repo_context.py --cwd <directory>The script prints one JSON document. It runs only read-only git queries that execute
no repository-configured command, follows no symlink at any path component, skips
dependency and build directories, and never runs the commands it lists.
Read the result
| Field | Meaning |
source | git (root, HEAD, branch, dirty counts) or archive-or-unversioned |
scan | Depth/directory budget, directories visited, depth-pruned and count-truncated |
instructions | Instruction and rule files with size and SHA-256; content is not copied |
ecosystems | Per-directory Node, Python, Go and Makefile declarations |
commandCandidates | argv + basis + executed: false; argv: null when blocked |
ci | CI files present; gates is always null because files are not interpreted |
unknowns | Fixed reason codes such as invalid-json, symlink-not-followed |
executionStillNeeded | What must still be run or confirmed |
Reconcile before acting
- Read every listed instruction file. When instructions and discovered facts disagree, the instructions win; report the conflict instead of silently choosing.
- Treat
defaultBranch: null,ci.gates: nullandpackageManager: nullas unknowns to confirm from instructions or the user. Do not infer them from lockfiles, conventions orlocalOriginHead. dirty: nullwithdirty-state-uninspected-content-filtersmeans a tracked path is routed through a configured filter driver; checking it would run that driver's command, so it was not checked. Establish the state by a means you trust.- Treat every candidate as unverified until it runs. Record runs with the
verified-handoffskill so the result is backed by a receipt. - For an
archive-or-unversionedsource, describe the source explicitly (name and digest) wherever a commit would otherwise be cited. - Resolve each entry in
unknownsbefore relying on the declaration it affects;directory-scan-depth-limitordirectory-scan-count-limitmeans declarations may exist beyond the scanned tree.
Limits
- Discovery is bounded (depth 4, 5000 directories, 1 MiB per file).
_test.gocounts cover only the scanned tree. Submodule state is not inspected. - Declared facts are not installed facts: toolchain versions and dependencies must still be confirmed in the target environment.
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
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.