MCPcopy Index your code
hub / github.com/cameroncooke/skills

github.com/cameroncooke/skills @main

Chat with this repo
repository ↗ · DeepWiki ↗ · + Follow
96 symbols 308 edges 7 files 0 documented · 0%
What it actually does AI analysis from the code graph — generated when you open this
loading…
README

Agent Skills

AXe

A curated collection of custom skills for AI coding agents — designed to automate repetitive developer workflows around releases, changelogs, code review, and knowledge sharing.

Skills

changelog-updater

Keeps your CHANGELOG up to date by analysing git history and writing user-facing release notes.

Compares commits between the latest tag and main, clusters them by outcome, and writes concise Added / Changed / Fixed / Breaking sections. Handles prerelease tags (e.g. v2.0.0-beta.1 merges into the canonical 2.0.0 section), attributes external contributors, and links new CLI commands or config options to their docs. Breaking changes include migration guides with before/after code examples.

Invoke with: /changelog-updater


pr-learning

Mines PR review feedback to surface repeatable rules and learnings for your team.

Collects review comments across one or more PRs using gh, normalises them into observations, and clusters repeated patterns. Proposes ranked Rule (strict) and Learning (soft) candidates with confidence scores, dedupes against items already in your AGENTS.md / CLAUDE.md, and — after your approval — codifies them with full provenance tracking (source PRs, semantic keys, confidence).

Invoke with: /pr-learning


release-tweet

Drafts a release announcement tweet from a GitHub release — ready for you to review and post.

Fetches release notes via gh, parses Added/Changed/Fixed/Removed sections, and condenses them into outcome-focused bullet points. Identifies external contributors, cross-references their X/Twitter handles with multi-signal verification, and outputs the composed tweet text alongside a confidence table for each handle lookup.

Invoke with: /release-tweet


agent-change-walkthrough

Generates a narrative walkthrough of AI-authored code changes, ordered by dependency then runtime flow.

Weaves changed and unchanged code into a single story — from the trigger that kicks things off to the final observable behaviour. Each step includes annotated mini-diffs with path/to/file:line headers, concrete input/output examples for data-shape changes, and inline notes on trade-offs, alternatives, and risks.

Invoke with: /agent-change-walkthrough


reconcile-merge-conflicts

Resolves merge and rebase conflicts with branch-aware reconciliation and regression-focused decision making.

Detects whether a merge or rebase is active, asks for the target when none is in progress, then resolves conflicts file-by-file by preserving important intent from both sides while treating the target branch (usually main) as canonical for current behavior. Produces a confidence-ranked reconciliation audit with risks, evidence, and open questions.

Invoke with: /reconcile-merge-conflicts


pr-comment-resolution

Audits and resolves PR review comments with an evidence-first workflow.

Collects PR feedback through the skill’s helper scripts, then classifies each item as valid/invalid/contentious/already-addressed/out-of-scope using direct code evidence. Audits inline review threads, general PR comments, and optional review summaries with the same evidence-first process, including broader question-style feedback that requires targeted code reading. After approval, it can help apply scoped fixes, post follow-up comments or replies as appropriate, and resolve threads with commit-linked rationale.

Invoke with: /pr-comment-resolution


slop-swatter

Removes AI slop from PR-level code changes before merge while preserving business behavior.

Orchestrates role-scoped subagents to map staged, unstaged, and committed-but-unmerged changes; read full files and complete diffs; reuse existing project patterns; remove defensive overengineering, premature abstractions, boilerplate types, and readability-only helper functions; and run quality checks before handoff.

Invoke with: /slop-swatter


Installation

Install individual skills directly from this repo with npx skills add:

npx skills add cameroncooke/cameroncooke-skills@agent-change-walkthrough
npx skills add cameroncooke/cameroncooke-skills@changelog-updater
npx skills add cameroncooke/cameroncooke-skills@pr-learning
npx skills add cameroncooke/cameroncooke-skills@release-tweet
npx skills add cameroncooke/cameroncooke-skills@reconcile-merge-conflicts
npx skills add cameroncooke/cameroncooke-skills@pr-comment-resolution
npx skills add cameroncooke/cameroncooke-skills@slop-swatter

You can also install manually by copying skill directories into your agent's skills folder:

Agent Skills path
Claude Code ~/.claude/skills/
Others ~/.agents/skills/

Each skill is self-contained — copy only the ones you need.

Core symbols most depended-on inside this repo

run_json
called by 6
pr-comment-resolution/scripts/collect_pr_feedback.py
now_iso
called by 6
pr-learning/scripts/common.py
write_json
called by 6
pr-learning/scripts/common.py
run
called by 5
pr-comment-resolution/scripts/collect_pr_feedback.py
git_stdout
called by 5
slop-swatter/scripts/effective_pr_state.py
read_json
called by 4
pr-learning/scripts/common.py
run
called by 3
pr-comment-resolution/scripts/apply_resolution_actions.py
require
called by 3
pr-comment-resolution/scripts/apply_resolution_actions.py

Shape

Function 96

Languages

Python100%

Modules by API surface

pr-comment-resolution/scripts/collect_pr_feedback.py23 symbols
pr-learning/scripts/common.py17 symbols
slop-swatter/scripts/effective_pr_state.py15 symbols
pr-learning/scripts/build_candidates.py15 symbols
pr-learning/scripts/codify_learnings.py13 symbols
pr-learning/scripts/collect_feedback.py7 symbols
pr-comment-resolution/scripts/apply_resolution_actions.py6 symbols

For agents

$ claude mcp add skills \
  -- python -m otcore.mcp_server <graph>

⬇ download graph artifact