source-inventory
Skill5 starsChanged 3 months agoMITView on GitHub
Maintain the source registry in sources/supported-stores.json — add a new third-party plugin store, retire one, refresh metadata, validate the schema. Use when proposing a registry change, after the weekly cron flags a source as unhealthy, or when a candidate store needs evaluation before adding.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- bingyue/fde-skillstheir own version, first seen Oct 13, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- bingyue/fde-skills★ 5same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- .github/copilot-instructions.mdCopilot instructions
- currency-auditSkill
- mall-self-curationSkill
- meditationSkill
- staleness-disciplineSkill
- store-evaluationSkill
The file
We split this skill into the rules it gives an agent
11 rules. About 0 are common, 5 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 84
Use the /add-source prompt.
- Distinctive: line 86
Run store-evaluation first (scorecard ≥ 7 required)
- Very specific: line 88
Add the entry to supported-stores.json (alphabetical by name)
- Very specific: line 89
Run node scripts/scan-sources.cjs locally with SOURCES_DIR pointing at a temp dir
- Very specific: line 90
Verify the new store produces a plausible catalog/stores/<name>.json
- Very specific: line 91
Commit [behaviour] add source: <name> and let Monday's cron produce the first full catalog refresh
- Distinctive: line 97
Run staleness-discipline to confirm a staleness signal fires
- Very specific: line 98
Add a // pruned-pending: YYYY-MM-DD reason: <reason> line above the entry and comment the entry out
- Distinctive: line 100
Wait one weekly cron cycle to verify the catalog refresh PR shows the expected removal cleanly
- Distinctive: line 101
Delete the commented-out entry in the next commit
- Distinctive: line 106
When reviewing a weekly catalog-refresh PR, sanity-check each store:
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Source Inventory
Maintain sources/supported-stores.json — the single source of truth for which upstream plugin repos the Mall scans.
The Mall does NOT keep persistent local clones of source repos. The weekly cron clones each registered source into a workflow temp directory ($SOURCES_DIR), scans it, scores it, and discards the clone. The registry file is the only persistent record.
When to Use
- Proposing a new source store (run store-evaluation first)
- Retiring a source store (run staleness-discipline first; use
/prune-source) - A weekly cron PR flags a source as unhealthy (broken remote, archived, license drift)
- Schema validation after editing
supported-stores.json
The registry: sources/supported-stores.json
{
"$schema": "./supported-stores.schema.json",
"schema_version": "2.0",
"stores": [
{
"name": "plugin-mall", // unique kebab-case; matches catalog/stores/<name>.json
"remote": "https://github.com/<org>/<repo>", // upstream URL (informational; not cloned for plugin-mall)
"pluginDir": "plugins", // where plugins live in the upstream repo
"quality": "first-party", // tier tag — see § Quality tiers
"provenance": true, // true ONLY for plugin-mall (drives +50 trust signal)
"license": "PolyForm-Noncommercial-1.0.0", // SPDX id, or null if unknown
"added_at": "2026-05-29"
},
{
"name": "awesome-copilot",
"remote": "https://github.com/github/awesome-copilot",
"pluginDir": "skills",
"quality": "official",
"provenance": false,
"license": "MIT",
"added_at": "2026-04-15"
}
]
}Required fields
| Field | Type | Notes |
name | string | Kebab-case, ≤32 chars, unique across the registry; produces catalog/stores/<name>.json |
remote | string | Upstream URL (HTTPS). For plugin-mall, this is informational; for everyone else, it's what bootstrap-sources.cjs clones |
pluginDir | string | Path under the upstream repo where plugins live (., plugins, skills, agents, etc.). Use . if plugins live at the repo root. |
quality | enum | See § Quality tiers below |
provenance | bool | true ONLY for plugin-mall (drives the +50 trust signal). false for all third-party stores. |
license | string or null | SPDX identifier (e.g., MIT, Apache-2.0, PolyForm-Noncommercial-1.0.0). null if unknown — license signal will score 0. |
added_at | string | YYYY-MM-DD of registry add |
Optional fields
| Field | Type | Notes |
local_dir_name | string | Override the directory name used inside $SOURCES_DIR when bootstrapping. Defaults to name. Use when the upstream repo has a different conventional folder name. |
note | string | Human-readable note (audit trail) |
pruned_at | string | YYYY-MM-DD of removal (use during the two-step prune; final delete removes the entry entirely) |
Quality tiers
| Tier | Meaning | Examples of fit |
first-party | The Mall's own curated plugins (plugin-mall only) | plugin-mall |
official | Maintained by a major platform vendor with editorial standards | awesome-copilot (GitHub), microsoft-skills |
community-curated | Community repo with explicit curation standards | awesome-mcp-servers, awesome-claude-code |
community | Individual or small-team plugin collections | personal skill repos |
domain | Vertical-specific plugin collections | game-dev, marketing, healthcare |
reference | High-signal individual contributors | named researchers, well-known authors |
The quality tier is informational — it does NOT directly affect the trust score. Trust is computed from the six published signals (provenance, maintenance, adoption, license, frontmatter, README). Quality tier helps reviewers triage; the signals do the scoring.
Adding a new source
Use the /add-source prompt. Sequence:
- Run store-evaluation first (scorecard ≥ 7 required)
- Determine
name,pluginDir,quality,license - Add the entry to
supported-stores.json(alphabetical byname) - Run
node scripts/scan-sources.cjslocally withSOURCES_DIRpointing at a temp dir - Verify the new store produces a plausible
catalog/stores/<name>.json - Commit
[behaviour] add source: <name>and let Monday's cron produce the first full catalog refresh
Retiring a source
Use the /prune-source prompt. Sequence:
- Run staleness-discipline to confirm a staleness signal fires
- Add a
// pruned-pending: YYYY-MM-DD reason: <reason>line above the entry and comment the entry out - Commit
[behaviour] prune source (pending): <name> - Wait one weekly cron cycle to verify the catalog refresh PR shows the expected removal cleanly
- Delete the commented-out entry in the next commit
- Commit
[behaviour] prune source (delete): <name>
Health checks (run during PR review)
When reviewing a weekly catalog-refresh PR, sanity-check each store:
| Check | Pass | Action on fail |
bootstrap-sources.cjs cloned the remote successfully | Workflow log shows clone success | Investigate remote; remote may have moved or repo may be deleted — surface in PR comments |
scan-sources.cjs produced a non-zero plugin count | Plugin count > 0 in catalog/stores/<name>.json | Check pluginDir — upstream may have renamed the plugins folder |
fetch-github-stats.cjs returned data | Stars / contributors / last_commit populated in scoring/github-stats.json | API rate-limit hit, OR repo is archived/private — surface in PR |
compute-trust.cjs produced a non-null score | trust_score populated in catalog/stores/<name>.json | A signal is missing or unparseable — check scoring/trust-audit.md for the per-store breakdown |
Anti-Patterns
| Anti-pattern | Correction |
| Adding a source without running store-evaluation | The scorecard is the bar; skip it and stale stores accumulate |
Setting provenance: true for a third-party store | Provenance is for plugin-mall only — it drives the +50 first-party trust signal, not "I trust this maintainer" |
Hand-editing catalog/stores/<name>.json | Catalog is generated; the only thing you edit is the registry |
| Deleting a registry entry in one commit | Always use the two-step prune so one cron cycle can verify the catalog refresh PR is clean |
| Skipping the schema validation step | The schema catches simple errors (typo'd field name, wrong enum value, malformed URL) before they break the workflow |
Falsifiability
This skill is wrong if any of the following occur by 2026-08-29 (90 days):
- A registry entry that passed
/add-sourcereview gets pruned within 30 days (scorecard bar too lax) - A pruned store gets re-added within 30 days (staleness signals miscalibrated)
- The schema lets through an invalid entry that breaks the workflow ≥1 time
- The quality-tier taxonomy needs revision ≥2 times in a quarter (taxonomy is wrong shape)
Track outcomes in docs/curation-log.md tagged [SOURCE-INVENTORY].
Related
- store-evaluation — gate before adding
- staleness-discipline — gate before pruning
- mall-self-curation — the pipeline this registry feeds
/add-sourceprompt/prune-sourcepromptsources/supported-stores.schema.json— JSON schema
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

