agentleFS
Sign inStart free
aiagentkit-labsaiagentkit-labs/agentic-dev-kit/skills/branch-policies/SKILL.md

branch-policies

Skill0 starsChanged 27 days agoMITView on GitHub

Review and set branch policies on Azure Repos or GitHub through an interview, including relaxing approval rules for a solo developer with no one to approve their PRs. Use when the user mentions branch policies, branch protection, required reviewers, or working alone on a repo.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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. Very specific:

    Run the preflight and postflight as the other skills do (check.sh pre|post branch-policies), and show each table.

    line 8
  2. Distinctive:

    When a second person joins, switch back to team.

    line 26
  3. Very specific:

    Azure Repos, solo: update the minimum-reviewers policy with az repos policy approver-count update --id <policy-id> --minimum-approver-count …

    line 43
  4. Very specific:

    Azure Repos, both profiles: create any missing policy the profile requires, for example az repos policy work-item-linking create and az …

    line 45

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Branch policies are shared infrastructure. Show every change as before and after, apply it only on the user's yes, and leave the user able to revert it. Changing policies needs admin rights: Edit policies on the Azure Repos repository, or admin on the GitHub repository.

Run the preflight and postflight as the other skills do (check.sh pre|post branch-policies), and show each table. Use the interview mode from implement, with each gate a round in that format.

Profiles

Policy on the default branchTeamSolo
Changes arrive through a pull requestrequiredrequired
Minimum approvers1+, the author's vote doesn't countAzure Repos: 1, the author's own vote counts. GitHub: 0 (GitHub never lets authors approve)
Reset approvals on new pushesonon
Build validation / required status checksrequiredrequired
Linked work item (Azure Repos only, if enabled by the repo)requiredrequired
Comments resolved before mergerequiredrequired
Force pushes and branch deletionblockedblocked

Solo relaxes only who approves. Everything that catches mistakes stays. In return, the kit adds controls:

  • The PR, Security, and Privacy reviewers act as the second reviewer. Their merged findings are posted on the PR as its review record.
  • An escalate finding has no sign-off team, so the user acknowledges each one in a PR comment before completing.
  • When a second person joins, switch back to team.

1. Read the current state

  • Azure Repos: get the repository ID with az repos show --repository <name> --query id -o tsv, then az repos policy list --repository-id <id> --branch <branch>.
  • GitHub: gh api repos/<owner>/<repo>/branches/<branch>/protection (a 404 means unprotected) and gh api repos/<owner>/<repo>/rulesets. Count who else could approve with gh api repos/<owner>/<repo>/collaborators --jq length.

Done when every policy on the branch is listed with its current values.

2. Choose the profile

Gate: recommend solo when the user is the only one who can approve, otherwise team. Show a table of each policy (current value, target value) and the exact commands from step 3. Write down the revert commands (the current values) to include in the report.

3. Apply

On the user's yes:

  • Azure Repos, solo: update the minimum-reviewers policy with az repos policy approver-count update --id <policy-id> --minimum-approver-count 1 --creator-vote-counts true --reset-on-source-push true. If none exists, create one with az repos policy approver-count create --repository-id <id> --branch <branch> --blocking true --enabled true --minimum-approver-count 1 --creator-vote-counts true --allow-downvotes false --reset-on-source-push true.
  • Azure Repos, team: the same command with --creator-vote-counts false and the team's approver count.
  • Azure Repos, both profiles: create any missing policy the profile requires, for example az repos policy work-item-linking create and az repos policy comment-required create, each with --repository-id <id> --branch <branch> --blocking true --enabled true. Build validation needs a pipeline ID (az repos policy build create --build-definition-id <id> ...); offer it only when a pipeline exists.
  • GitHub, branch protection: gh api -X PATCH repos/<owner>/<repo>/branches/<branch>/protection/required_pull_request_reviews -F required_approving_review_count=<0 for solo, 1+ for team>.
  • GitHub, rulesets: fetch the ruleset, change required_approving_review_count in its pull_request rule, and gh api -X PUT repos/<owner>/<repo>/rulesets/<id> --input .git/agentic-dev-kit/tmp/ruleset.json.

Branch policies are never bypassed and never disabled wholesale. The guardrail hook asks before every policy change.

4. Verify and report

Re-read the policies (step 1). Done when they match the target. Report before and after, the revert commands, and what the profile means day to day: in solo, implement posts the reviewers' findings on each PR and asks before the user approves their own.

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.

Check your setup, free