Discover skills
Skill LaWebcapsule/d9-skills/skills/skillops/discover-skills
Agent Skills for d9 (Directus 9 fork) — install with: npx skills add LaWebcapsule/d9-skills
npx -y skills add LaWebcapsule/d9-skills --skill discover-skillsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Identify missing skills and recommend installations from local inventory or public skill catalogs.
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
5.2 KB, as published. Nobody here has run it
Skill: Discover Skills
Purpose
Identify capability gaps in the current skill inventory and recommend 1-3 skill installations from known catalogs (local repo, ai-cortex, public skill registries). This helps match-existing find skills beyond the local project and helps contributors avoid creating skills that already exist elsewhere.
Core Objective
Primary Goal: Provide the agent with top 1-3 skill recommendations and exact installation commands to fill capability gaps.
Success Criteria:
- Capability gap is clearly identified from the task context
- Local skills are checked first before searching external catalogs
- Each recommendation includes rationale (why this skill fits the gap)
- Install commands are provided and ready to use
- No installation happens without explicit user confirmation
Acceptance Test: The user receives actionable recommendations they can install in under 1 minute.
Use Cases
- During match-existing: extend the search beyond local skills to find cross-repo duplicates
- New project setup: identify which skills from the ecosystem would benefit the project
- Gap analysis: after a session with many errors, check if existing skills could have prevented them
- Catalog browsing: user asks "what skills exist for Docker?" or "are there skills for AWS?"
Behavior
Discovery Flow
- Understand the gap: analyze the task description or error pattern
- Search local first: check
.claude/skills/for matching skills - Search known catalogs:
nesnilnehc/ai-cortex(33 skills, meta/review/workflow)skills.shdirectory (if accessible)- GitHub search for repos with
.claude/skills/orSKILL.md
- Rank matches: by relevance to the gap, ASQM score, and freshness
- Present recommendations:
I found 2 skills that could help: 1. review-security (ai-cortex, ASQM 19) Covers: SQL injection, XSS, auth bypass patterns Install: npx skills add nesnilnehc/ai-cortex/review-security 2. docker-hardening (community, ASQM 14) Covers: Dockerfile best practices, multi-stage builds Install: npx skills add user/repo/docker-hardening Install one of these? [1] [2] [Skip] - Confirm before installing: never auto-install
Integration with match-existing
When called by match-existing, the flow is:
- Receive the extracted experience from detect-xp
- Search external catalogs for similar skills
- Return matches with similarity assessment
- match-existing combines local + external results for deduplication
Input & Output
Input: Task description, error pattern, or capability gap query. Optionally: list of currently installed skills.
Output:
- List of 1-3 recommended skills with: name, source repo, ASQM score (if available), description, install command
- Gap analysis: what the current inventory is missing
- Deduplication note: if the user is about to create something that already exists elsewhere
Restrictions
Hard boundaries
- Do NOT install skills without explicit user confirmation
- Do NOT recommend more than 3 skills at once (cognitive overload)
- Do NOT recommend skills with ASQM < 10 (quality floor)
- Do NOT search repositories the user hasn't authorized
Skill boundaries
- If a skill needs to be created → hand off to
format-skill - If a found skill needs quality assessment → hand off to
curate-skills - If a found skill needs adaptation → hand off to
refine-skill-design
Self-Check
- Local skills were checked before external catalogs?
- Each recommendation has a clear rationale?
- Install commands are correct and ready to use?
- No more than 3 recommendations presented?
- User was asked for confirmation before any installation?
- Deduplication risk was flagged if relevant?
Examples
Example 1: Gap detection during work
Context: Developer is setting up AWS infrastructure and makes a region configuration mistake.
Discovery:
I noticed you're configuring AWS services. Your project already has
`ses-eu-west3-setup` but you might benefit from:
1. aws-region-consistency (ai-cortex, ASQM 16)
Ensures all AWS services use the same region.
Install: npx skills add nesnilnehc/ai-cortex/aws-region-consistency
Want to install this? [Yes] [No]
Example 2 (edge case): No match found
Context: Developer asks about skills for an obscure framework.
Discovery:
I searched local skills and known catalogs — no skills found for
[framework]. This could be a good opportunity to create one from
your experience. Want to start the contribution pipeline?
[Yes, create a skill] [No]