project-profile
Skill0 starsChanged 25 days agoMITView on GitHub
Load, discover or update the current project's dev-skills profile: issue tracker, PR host, base branch, CI, test/lint/boot commands, staging and production deploy, feature flags, observability, analytics, console access, API docs and project-specific skills, plus the user's own notes. Every dev-skills skill runs this first. Also use when the user says "remember for this project: ...", "show project profile", "refresh profile", or a profile fact turns out to be wrong.
- Reads credentials
6 other projects use a version of this.
Adopted by
Most recently taken up by
- brilliantrough/agent-skillstheir own version, first seen Nov 8, 2026
- Zejnilovic/agdtheir own version, first seen Nov 1, 2026
- juanitourquiza/shipframetheir own version, first seen Oct 23, 2026
- awahids/noskilltheir own version, first seen Oct 15, 2026
- analyticsengineering/ae-toolkittheir own version, first seen Oct 12, 2026
- jlrrequina/pi-project-profiletheir own version, first seen Sep 29, 2026
Used with
Projects with it also use tdd.
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.
GoogleSkilltdd
Test-driven development. Use when the user wants to build features or fix bugs test-first, mentions "red-green-refactor", or wants integration tests.
Also in 3 projects that use this
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- brilliantrough/agent-skills★ 11same name
- Zejnilovic/agd★ 5same name
- juanitourquiza/shipframe★ 4same name
- analyticsengineering/ae-toolkit★ 3same name
- awahids/noskill★ 1same name
- jlrrequina/pi-project-profile★ 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 repository18 other files it gives its agents.
- address-commentsSkill
- demoSkill
- deploySkill
- documentSkill
- gather-contextSkill
- implementSkill
- investigateSkill
- observeSkill
- pr-follow-upSkill
- pr-reviewSkill
- qa-planSkill
- qaSkill
- resolve-conflictsSkill
- specSkill
- sprintSkill
- standupSkill
- todoSkill
- triageSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 13
Follow ../../shared/policy.md (relative to this skill's folder).
- Very specific: line 21
Exists: read it. Done, unless the caller needs a field that is unknown: discover just that field (step 3) and update the file.
- Very specific: line 24
Do not ask the user to confirm anything: inferred facts are marked inferred.
- Very specific: line 48
Leave a field unknown rather than guessing without a signal.
- Distinctive: line 77
When a skill's command from the profile fails (wrong test command, missing tool), rediscover that field, update it, and mention the change …
“Common” is measured: kept by at least 6 in 10 of the 5 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
Project profile
Follow ../../shared/policy.md (relative to this skill's folder).
1. Locate
Compute <project-slug> (see policy "Paths") for the repo at the current working directory, or
the repo the task is about. The profile is ~/.dev-skills/profiles/<project-slug>.md.
Worktrees of the same repo share the slug, so they share the profile.
2. Load or discover
- Exists: read it. Done, unless the caller needs a field that is
unknown: discover just that field (step 3) and update the file. - Missing: discover everything (step 3), write the file (step 4), and print a 5-10 line summary.
Do not ask the user to confirm anything: inferred facts are marked
inferred.
3. Discover
Reads only. Record each value with its source (file:line, command, or user) and confidence
(confirmed when read directly from config or code, inferred when guessed from signals).
| Field | Signals |
pr_host | git remote get-url origin host; tool available (gh, glab, Bitbucket MCP via ToolSearch) |
tracker, ticket_key | ticket keys in git branch -a and recent git log subjects; .github/ISSUE_TEMPLATE; links in README/CONTRIBUTING; connected tracker MCPs (ToolSearch "jira", "linear", "clickup"); app.clickup.com links |
base_branch, branch_naming, pr_template | git symbolic-ref refs/remotes/origin/HEAD; recent branch names; .github/pull_request_template.md or similar; docs |
ci | .github/workflows, bitbucket-pipelines.yml, .gitlab-ci.yml, .circleci |
commands.test, commands.lint, commands.boot, commands.setup | repo CLAUDE.md, README, CONTRIBUTING, docs/, package.json scripts, Makefile, Procfile, compose files, pre-commit config |
env_files | .env.example/.env.sample and which untracked env files the docs say to create |
deploy.staging, deploy.prod, deploy.order | see ../../shared/references/deploy.md "Finding the method"; docs; deploy scripts |
flags | dependency files (see deploy.md "Feature flags") |
errors, logs, metrics, traces, apm | dependencies (sentry, rollbar, honeybadger, bugsnag, ddtrace, newrelic, opentelemetry, prometheus client); connected MCPs (ToolSearch "grafana", "sentry", "datadog"), mapped per environment |
analytics | amplitude, mixpanel, segment, posthog, statsig, growthbook dependencies |
console | Gemfile → Rails, manage.py → Django, ORM REPL; how prod console is reached if documented |
api_docs | openapi.*, swagger.*, Postman collections (*.postman_collection.json), docs/api, API sections in README |
docs_tool, chat_tool | connected Notion/Confluence/Google Drive, Slack MCPs |
project_skills | .claude/skills/*/SKILL.md and .claude/commands/*.md in the repo: name + one-line purpose, and which dev-skills step each replaces (e.g. a staging deploy skill → deploy staging) |
related_repos | sibling repos with profiles that share the same tracker board; frontend/backend pairs under one parent folder |
Leave a field unknown rather than guessing without a signal.
4. File format
# <project-slug>
Repo: <absolute path of main checkout> · Updated: <YYYY-MM-DD>
## Discovered
<!-- Rewritten by project-profile. Edit ## Mine instead. -->
- pr_host: github (gh) — git remote · confirmed
- tracker: jira, key DEV-\d+ — branch names · inferred
- commands.test: bin/rails test — README.md:42 · confirmed
- deploy.staging: project skill deploy-to-staging — .claude/skills · confirmed
- metrics: grafana-prod (prod), grafana-dev (staging) — connected MCPs · inferred
- analytics: unknown
## Mine
<!-- Yours. Never rewritten. A key here overrides the same key above. -->
### general
### implement
### deploy
### standup
### planRewrite only ## Discovered. Never change ## Mine except to append a note the user asked to
remember (policy "Project config") or a mode: line set by /dev-skills:mode ... project.
5. Keep it right
When a skill's command from the profile fails (wrong test command, missing tool), rediscover that field, update it, and mention the change in one line.
Output
For the caller: the facts it needs. For the user: the summary only when the profile was created or changed. When the user asked to see it, print the file.
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.

