Agent skills for working with AppMap runtime recordings. Each
directory is a self-contained skill (a SKILL.md the agent reads, plus any assets).
| Skill | What it does |
|---|---|
appmap-record |
Record runtime data from an application with the AppMap agents. |
appmap-config |
Configure what gets recorded: appmap.yml (packages, exclude, shallow, output dir) and function labels, per language. |
appmap-gold-traces |
Maintain a committed baseline of curated recordings ("gold traces") and bless it as code evolves. |
appmap-review |
Diff a change's runtime behavior against the baseline and write an interpreted code review. |
appmap-setup |
Get a repository recording AppMap data: config, one unit and one integration recording, noise exclusions, and a docs/appmap.md imported from CLAUDE.md with the working record commands. |
appmap-setup-review |
Demonstrate a gold-trace review on an existing change: set up recording and gold traces on a BASE commit, replay onto HEAD, and run appmap-review between them. |
Symlink (or copy) the skills you want into ~/.claude/skills, then invoke them by name:
ln -s "$PWD/appmap-review" ~/.claude/skills/appmap-reviewThe getappmap/review-action GitHub
Action installs these skills and runs appmap-gold-traces + appmap-review on a pull
request.