codex-ab
Run an A/B codex review experiment — holistic codex review vs 3 focused dimension passes (security, ecto, liveview) on the branch diff, classify findings,…
CONTRIBUTOR TOOL - Track CC changelog, extract new versions since last check, analyze impact on plugin (breaking changes, opportunities, deprecations). Run periodically or before releases. NOT part of the distributed plugin.
$ npx -y skills add oliver-kriska/claude-elixir-phoenix --skill cc-changelog --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/cc-changelogContext preview
The summary Claude sees to decide when to auto-load this skill.
CONTRIBUTOR TOOL - Track CC changelog, extract new versions since last check, analyze impact on plugin (breaking changes, opportunities, deprecations). Run periodically or before releases. NOT part of the distributed plugin.
name: cc-changelog description: | CONTRIBUTOR TOOL - Track CC changelog, extract new versions since last check, analyze impact on plugin (breaking changes, opportunities, deprecations). Run periodically or before releases. NOT part of the distributed plugin. argument-hint: "[--full|--set=VERSION]" effort: medium
Tracks Claude Code releases against the plugin. Fetches the CC changelog, extracts entries newer than last check, and analyzes impact on plugin components (agents, skills, hooks, config).
/cc-changelog # Check for new CC versions, analyze impact /cc-changelog --full # Re-analyze all versions (ignore last check) /cc-changelog --set=2.1.85 # Reset last checked version (then re-run)
bash "${CLAUDE_PROJECT_DIR}/scripts/fetch-cc-changelog.sh"If output starts with `STATUS: UP_TO_DATE` → report "No new CC versions" and stop.
If `STATUS: NEW_VERSIONS` → continue with the changelog content below the header.
For `--full`: run `bash "${CLAUDE_PROJECT_DIR}/scripts/fetch-cc-changelog.sh" --all` For `--set=X`: run `bash "${CLAUDE_PROJECT_DIR}/scripts/fetch-cc-changelog.sh" --set=X`, then re-run without flag.
Read the new changelog entries. For EACH entry, classify into one of:
| Category | Meaning | Action | |----------|---------|--------| | **BREAKING** | May break existing plugin functionality | Immediate fix required | | **OPPORTUNITY** | New CC feature the plugin could use | Add to backlog/plan | | **RELEVANT FIX** | CC fixed a bug we worked around | Check if workaround can be removed | | **DEPRECATION** | CC removing something we use | Plan migration | | **INFO** | Good to know, no action needed | Log in memory update |
Cross-reference against plugin components using rules in `${CLAUDE_SKILL_DIR}/references/analysis-rules.md`.
Output a structured report:
## CC Changelog Analysis: v{last_checked} → v{latest}
### BREAKING (action required)
- [version] description → **Impact**: which plugin file/component
**Fix**: specific action needed
### OPPORTUNITY (new features)
- [version] description → **Use case**: how plugin could benefit
**Files**: which plugin files to update
### RELEVANT FIX (workaround removal)
- [version] description → **Current workaround**: what we do now
**Action**: can we simplify?
### DEPRECATION (migration needed)
- [version] description → **We use this in**: file:line
**Migration**: what to change
### INFO (no action)
- [version] brief summary (collapsed)After user reviews the report, ask:
> "Update last checked version to {latest}? This also updates the > CC internals memory file. [Yes/No]"
If yes: 1. Run `bash "${CLAUDE_PROJECT_DIR}/scripts/fetch-cc-changelog.sh" --set={latest}` 2. Update memory file `reference_cc_source_internals.md` with new findings 3. If BREAKING or DEPRECATION items found, offer to create a plan
1. **ALWAYS fetch before analyzing** — never analyze stale cache 2. **NEVER auto-update state** — user must confirm after reviewing report 3. **ALWAYS cross-reference plugin files** — don't just summarize, map to impact 4. **BREAKING changes are BLOCKERS** — surface first, prominently 5. **Track the audit version** — state file is the source of truth
Docs: phxagents.dev -- install guides per runtime, the runtime compatibility matrix, all 26 Iron Laws, and a browsable skill and agent catalog. Claude Code is great.
Repo: oliver-kriska/claude-elixir-phoenix
Run an A/B codex review experiment — holistic codex review vs 3 focused dimension passes (security, ecto, liveview) on the branch diff, classify findings,…
CONTRIBUTOR TOOL - Validate plugin against latest Claude Code documentation. Catches breaking changes, deprecations, discovers new features. Run before…
Guide plugin development workflow — editing skills, agents, hooks, or eval framework in this repo. Use when modifying files in plugins/elixir-phoenix/,…
Generate X/Twitter release promotion posts with ASCII tables and CodeSnap rendering. Use when writing release posts, promotion tweets, plugin announcements, or…
CONTRIBUTOR TOOL - Cut a plugin release: bump plugin.json version, finalize CHANGELOG, update README if needed, gate on make ci, commit, tag vX.Y.Z, and create…
Deep qualitative analysis of high-signal sessions. Spawns subagents with v2 template, synthesizes patterns, compares against known findings. Use after…