list-skills
Table of Contents generated with DocToc
Pre-flight — is this project set up?
Do this first, before anything else in this skill, and do it silently: on the happy path it costs three file checks and prints nothing.
A marketplace install delivers skills only. Nothing in it configures this
repository, and on most harnesses no code runs at all when a plugin is
installed or upgraded — there is no post-install step to rely on. Claude Code's
SessionStart hook covers only the all-in-one plugin, so for every other
install this check is the one thing standing between a stale or unadopted repo
and a skill that acts on wrong assumptions.
- Is a lock present? If
.apache-magpie.lockexists, this project uses the pinned-snapshot install. Compare it with.apache-magpie.local.lock:- local lock missing → the snapshot was never fetched on this machine;
ref/commitdiffer → this machine is on a different framework version than the project pins.
- No lock? Then this is the marketplace install (or nothing at all). Look
for a
<project-config>/directory. If there is none, the project has not been adopted and every<placeholder>in this skill is unresolved. - Anything unresolved above → stop and propose
/magpie-setup(or/magpie-setup upgradefor a version mismatch). Say which of the three checks failed and what you found. Do not run setup unattended and do not continue this skill on a guess: a skill that proceeds against an unadopted repo writes to the wrong tracker.
Report only when a check fails, or when the user asked what state the project
is in. /magpie-setup verify is the full diagnostic — this is deliberately the
cheap subset that is worth paying for on every invocation.
Print a human-readable index of the skills installed for this
repository. The index is generated on every run from live
SKILL.md frontmatter — there is no cached copy to keep in sync.
The skill exists for humans (newcomers reading the repo,
maintainers checking what is available); agents route invocations
via the same frontmatter the script reads, so this skill is
purely informational.
What counts as "installed" depends on how Magpie was put in place, so the script covers all three shapes and labels which one each entry came from:
| Install | Where the skills live | Invocation shown |
|---|---|---|
| Pinned snapshot | .agents/skills/ plus the per-agent relays beside it |
/magpie-<skill> |
| Framework checkout | the repo's own skills/ |
/magpie-<skill> |
| Marketplace plugin | the plugin cache this script runs from, and its sibling plugins | /<plugin>:<skill> |
Prerequisites
- Python 3.11+ on
PATH. Nothing else — the script is stdlib-only, asskills/pyproject.tomlrequires of every helper script in this tree, and declares that contract in PEP 723 inline metadata.uv run --scriptand a barepython3therefore behave identically.
Step 1 — Run the listing script
Run the bundled script and present its output to the user verbatim:
python3 .claude/skills/magpie-list-skills/scripts/list_skills.py
Run that command literally, as written — do not expand it to
an absolute path. It is a repository-relative path that resolves
under both install methods that put skills in the repository: a
pinned snapshot install and the framework checkout both carry
.claude/skills/magpie-list-skills as a symlink onto the real
skill directory.
For a layout that puts each description on its own indented line
(easier to read when descriptions are long), pass --verbose; to
inspect a repository other than the enclosing one, pass --root:
python3 .claude/skills/magpie-list-skills/scripts/list_skills.py --verbose
python3 .claude/skills/magpie-list-skills/scripts/list_skills.py --root /path/to/repo
Marketplace installs are the one exception. They write nothing into the repository, so that path does not exist — the skill lives in the plugin cache. Build the command from the base directory reported for this skill instead:
python3 <the base directory reported for this skill>/scripts/list_skills.py
The script:
- resolves the repository from
git rev-parse --show-toplevel(or--root), not from its own location — under a per-family plugin install its own location is one family, not the whole install; - walks the agent-target directories that install writes into
(
.agents/skills/and its relays — the registry in../setup/agents.mdis the source of truth), the framework's ownskills/when the repo is the framework checkout, and the sibling plugins in the marketplace cache when it is running from one; - de-duplicates by the name you would type, so relay directories collapse to one entry while a skill available from two install methods keeps both — they are two different things to type;
- groups by each skill's declared
family:frontmatter key, per Golden rule 8. Family is never inferred from the name prefix:repo-healthandcontributor-growthspan several prefixes, andwrite-skillis familyutilities, not familywrite. A skill that declares no family lands inother; - prints each entry with the first sentence of its description, then a summary of which install each entry came from.
Step 2 — Hand the output to the user
Quote the script output back to the user as-is. Do not
paraphrase, summarise, or re-order — the value of this skill is
that the listing is the canonical, deterministic view of what
exists. If the user asks for more detail on a specific skill,
read that skill's SKILL.md and answer from it.
Hard rules
- Read-only. This skill never edits, creates, or deletes
files. It only reads
SKILL.mdfiles under the install directories listed above. - No paraphrasing. Always present the script output verbatim. Paraphrasing reintroduces the staleness this skill exists to prevent.
References
scripts/list_skills.py— the listing script Step 1 invokes.AGENTS.md— the framework's "Reusable skills" section, which explains the skills layout and frontmatter convention.../setup/agents.md— the agent-target registry the discovery list mirrors.../../docs/setup/marketplaces.md— why the invocation name differs between install methods.write-skill— sibling skill for authoring a new skill. Use it when the listing reveals a gap that warrants a new entry.