Agent skill · Audited from source

Finishing A Development Branch

skill · GanyuanRan/AegisMIT licenceAudited 18 Oct 2026 · source 177240a262fe

★ 1,357pushed today
✓SAFEGrade B · 89/100 · 18 Oct 2026 · 177240a
1,357repo stars · whole GanyuanRan/Aegis

Is it safe? Yes — SAFE, grade B (89/100). We read this skill's source at commit 177240a262fe and found no filesystem, network or shell access and no hard-coded secrets. What it does: Use when verified work needs integration or cleanup of an existing task-created branch/worktree, or the user explicitly requests merge, PR, or branch lifecycle handling.

Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
More
Cite

Cite this page

ModelsAtlas. “Finishing A Development Branch (GanyuanRan/Aegis)”. Retrieved 18 Oct 2026. https://modelsatlas.com/skills/ganyuanran/finishing-a-development-branch

6 skills are called “Finishing A Development Branch”. You are viewing GanyuanRan/Aegis.

and 1 more with this name.

Trust audit

Audit 0.4.1 · source read 18 Oct 2026

Trust score89/100✓SAFE B
0 critical0 high0 medium0 low0 info

SAFE means nothing in the source contradicts what the skill says it does. Grade B is shared by 13,985 of 14,816 audited skills.

What the source does

  • FilesystemNone observedNo reads or writes in the source
  • NetworkNone observedNo outbound calls in the source
  • ShellNone observedNo commands run in the source
  • DependenciesPinnedNothing floats
  • SecretsNone foundNo keys in source

Five audit layers

LayerWhat we checkResultFindings
L0 ProvenanceRepo owner, licence, archive status, activityPASS0
L1 Static sourceSecrets, path traversal, supply chain, hidden codeNA0
L2 InstructionsPrompt injection, authority framing, hidden directivesPASS0
L3 InventoryThe files and declarations match the sourcePASS0
L4 SandboxRun in isolation and watch behaviourSKIPPED why0

Gates applied: no behavioural pass.

Findings (0)

No findings at any severity. We re-audit when the source changes.

Audit 0.4.1 · source sha 177240a262fe · 18 Oct 2026Report an issue or request a re-scan

Provenance

Who published it, and what the verdict is pinned to

  • PublisherGanyuanRan GitHub userOwner of the repository this skill is read from
  • Source folderGanyuanRan/Aegis/skills/finishing-a-development-branchThe exact folder we audited
  • LicenceMITAs the repository states it
  • Commit pinned for this verdict177240a262feA new audit runs when the source changes
  • Repository statusNot archivedPushed today
  • Sandboxed runNot runSkills are read, not executed; layer L4 is skipped

What it tells the agent

Full source on GitHub ↗

Its description: Use when verified work needs integration or cleanup of an existing task-created branch/worktree, or the user explicitly requests merge, PR, or branch lifecycle handling.

---
name: finishing-a-development-branch
description: "Use when verified work needs integration or cleanup of an existing task-created branch/worktree, or the user explicitly requests merge, PR, or branch lifecycle handling."
---

# Finishing a Development Branch

## Overview

Integration choice and checkout lifetime are separate. Verify first, mutate only
task-owned resources, then prove what was removed or retained.

**Core principle:** fresh evidence -> authorized integration -> ownership-aware
cleanup -> Git/path readback.

**Announce at start:** explain which branch/worktree lifecycle is being closed.

## Step 1: Environment and Ownership Detection

Read the task's `TaskStartSnapshot` and current repository state. Record:

- root, `HEAD`, branch/detached state, upstream divergence, and active Git ops;
- staged/unstaged/untracked paths and the task-owned delta;
- `git worktree list --porcelain`;
- branch/worktree creator, task ownership, base branch, and cleanup trigger.

Stop on conflicts, unexplained detached state, failed verification, dirty/untracked state in
a cleanup target, or unknown ownership. Do not infer the base from the names
`main` or `master`; use task records, repository authority, or explicit user
direction. Never auto-stash, reset, clean, pull, rebase, amend, force-push, or
bypass hooks.

For a trusted host-managed detached checkout, follow the repository's approved
managed-workspace contract: verify host binding, default cwd, root, and HEAD;
then use its scoped local-commit and native handoff path. Report detached HEAD
in the receipt. This exception does not authorize automatic deletion of the
detached checkout or bypass unknown ownership/binding.

If no task-created branch/worktree exists and integration was not requested,
skip this workflow: report the local task commit plus `Task clean` and
`Repository clean`; do not invent merge/PR ceremony.

## Step 2: Fresh Verification

Run the smallest current-authority test set that proves the completed task. If
it fails, preserve all work and return to diagnosis. Do not present the branch
as ready.

## Step 3: Choose the Authorized Outcome

When a task branch actually needs disposition, present only applicable choices:

1. merge into the recorded base locally;
2. push and create/update a PR;
3. keep the branch for later integration;
4. discard the exact task-owned branch/worktree.

Push, PR, remote deletion, release, and force operations require explicit user
authorization. Discard requires a fresh, exact typed confirmation naming branch,
worktree path, and commits/delta to lose. Generic agreement is insufficient.

## Step 4: Execute Without Hidden Mutation

### Local merge

Verify the target checkout is safe, switch to the recorded base, merge without an implicit pull,
and rerun verification on the merged result. If integration
or verification fails, preserve both histories and stop.

### Push or PR

Push only the authorized task branch and create/update the PR with summary and
fresh test evidence. PR creation does not itself prove merge and does not
require the temporary worktree to remain.

### Keep branch

Retain the branch. Remove a clean task-created temporary worktree by default;
retain it only for an explicit ongoing-checkout need and report why.

### Discard

After exact confirmation, remove only the named clean task-owned worktree, then
the named branch. If either differs from the confirmation, stop and ask again.

## Step 5: Prove Integration Before Branch Deletion

Use evidence matching the merge strategy:

- merge/fast-forward: prove the task tip is an ancestor of the recorded base;
- squash/rebase: use fresh PR merge metadata plus strategy-appropriate patch
  equivalence (`git cherry` for rebased commits; stable patch-id or bounded
  changed-path/tree comparison for squashes), not an ancestor-only test;
- unmerged/open PR: keep the branch unless discard was exactly confirmed.

Never delete a branch merely because a PR was opened or a similarly named
commit exists.

## Step 6: Cleanup in Dependency Order

For a clean task-owned resource:

1. move to another safe registered checkout; never remove the current working directory;
2. remove/unregister the worktree first with `git worktree remove <exact-path>`;
3. read back `git worktree list --porcelain`;
4. verify the exact path no longer exists;
5. only then delete an integrated or exactly confirmed branch;
6. read back branch refs, `HEAD`, and repository status.

Do not use routine `--force`, global `git worktree prune`, wildcard deletion, or
cleanup of resources absent from the task ownership record. On Windows, Git may
unregister a worktree while its directory remains locked: re-check registration
and the exact resolved path, remove only a proven task-owned clean residual when
safe, otherwise retain it and report the lock/path. Never widen deletion scope.

## Final Git Receipt

Report:

- current branch and commit SHA/message, or the reason no commit exists;
- `Task clean: yes|no` and `Repository clean: yes|no` independently;
- branch/worktree created, removed, or retained, with evidence/reason;
- integration evidence and any remaining user action.

`Task clean` means no uncommitted task-owned delta remains. `Repository clean`
means no staged, unstaged, or untracked repository state remains.

This is a local Git receipt, not authoritative completion or merge truth beyond
the evidence read back.

## Integration

Called only when branch/worktree integration or explicit lifecycle handling is
in scope. Pairs with `using-git-worktrees` and
`verification-before-completion`.

Excerpt, first 5,601 characters of SKILL.md, verbatim from the audited commit. It is text from someone else’s repository, quoted so you can judge it before your agent reads it.

Install

Pinned to the audited commit 177240a

Download folder ↗
git clone https://github.com/GanyuanRan/Aegis
git -C Aegis checkout 177240a262fe

The author documents no install command, so this uses the repository directly. Our verdict covers commit 177240a262fe only.

Works in

The author has not said which agents this skill supports, and we have not tested it in any.

Audit history

Every audit, every source change

  1. ✓SAFE B 89 / 100

    first audit · source 177240a262fe

README badge

For maintainers · shows the latest audit

ModelsAtlas: SAFE BMODELSATLASSAFE · B
[![ModelsAtlas: SAFE B](https://modelsatlas.com/skills/ganyuanran/finishing-a-development-branch/badge.svg)](https://modelsatlas.com/skills/ganyuanran/finishing-a-development-branch/)

The badge shows our verdict and grade for this skill, and links back to this page. It changes when the audit does.

Brainstorming✓SAFE

GanyuanRan/Aegis

Use when defining ambiguous or high-complexity new features, product behavior, UI/component design, architecture choices, contract changes, or when grilling/pressure-testing a plan or design. Routine small requests stay…

★ 1.4kGrade B
Communicating Concisely✓SAFE

GanyuanRan/Aegis

Use when the user asks for caveman mode, fewer tokens, brief responses, compressed communication, or otherwise explicitly requests a much shorter answer.

★ 1.4kGrade B
Dispatching Parallel Agents✓SAFE

GanyuanRan/Aegis

Use when facing 2+ independent tasks without a written plan, with no conflicting shared mutable state or sequential dependencies, where parallel delegation beats inline cost; otherwise inline. Planned tasks use…

★ 1.4kGrade B
Establishing Project Context✓SAFE

GanyuanRan/Aegis

Use when the user asks to establish shared project language, or project work exposes a conflicting, renamed, or deprecated domain term that needs active semantic modeling. Routine small tasks stay on the fast path.

★ 1.4kGrade B
Executing Plans✓SAFE

GanyuanRan/Aegis

Use when executing a written implementation plan across sessions or with review checkpoints. Small or single-slice plans stay inline. For same-session independent tasks, use subagent-driven-development instead.

★ 1.4kGrade B
First Principles Review✓SAFE

GanyuanRan/Aegis

Use when asked for first-principles or Occam's-razor review, or when high-risk decisions involve competing constraints, fallback growth, duplicate owners, or architecture direction risk. Ordinary bug fixes stay on the…

★ 1.4kGrade B

Add this badge to your README

[![ModelsAtlas: SAFE B](https://modelsatlas.com/skills/ganyuanran/finishing-a-development-branch/badge.svg)](https://modelsatlas.com/skills/ganyuanran/finishing-a-development-branch/)

Questions about Finishing A Development Branch

What does the Finishing A Development Branch skill do?
Make AI coding agents architecture-aware: baseline-first, evidence-verified, drift-checked, and safe across long tasks.
Is Finishing A Development Branch safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Finishing A Development Branch access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
How current is this page?
The grade is for one exact copy of the source (177240a262fe), read on 2026-10-18. The repository is watched, and a new audit runs when it changes — this is the first audit.

Sources

Spotted something wrong? Report a correction · Methodology

Keep going in your AI

Opens with this page as the source. We add nothing else to the prompt.