AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Skill Index

skill-niborracho-skill-index-skill-index · by NiBorracho

Use when deciding which of 50+ installed skills to invoke. Reads a compact one-line-per-skill index to identify ALL relevant skills in one operation instead of invoking each individually.

No reviews yet
0 installs
39 views
0.0% view→install

Install

$ agentstack add skill-niborracho-skill-index-skill-index

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-niborracho-skill-index-skill-index)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
5mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Skill Index? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Skill Index

Overview

Problem: With 150+ installed skills, finding the right one by invoking each individually takes ~12.5 minutes (150 × 5s/call). The Skill tool loads full content even when you only need the 1-2 line description.

Solution: A compact INDEX.md file (~15KB for 150 skills) that Claude reads in one Read tool call, then identifies and invokes all relevant skills — no limit.

Impact: ~12.5 min → ~15 seconds to identify and load the right skills.


When to Use

| Situation | Action | |-----------|--------| | Choosing which skills to invoke (50+ installed) | Read INDEX.md first | | After installing a new skill | Run build-index.sh update | | After removing a skill | Run build-index.sh remove | | Index missing or corrupted | Run build-index.sh build | | Debugging skill discovery | Run build-index.sh debug |

Do NOT invoke skills blindly before reading the index when you have 50+ skills available.


Querying the Index

1. Read("~/.claude/skills/skill-index/INDEX.md")
2. Scan for entries matching keywords from the current task
3. Identify ALL relevant skills (no maximum — invoke every one that applies)
4. Skip entries marked as ~~deprecated~~
5. Invoke identified skills with the Skill tool

Reading the index costs 1 Read operation. Invoke as many skills as the task requires.


Index File Format

Location: ~/.claude/skills/skill-index/INDEX.md

Each skill entry uses HTML comment markers for reliable, auditable parsing:


- **skill-name** [source@version] — description from frontmatter
  • Markers guarantee idempotent updates (each add/remove targets exactly one block)
  • SHA256 detects when a skill file changes between sessions
  • Deprecated entries use strikethrough: ~~**skill-name**~~
  • Sections group skills by source plugin for readability

The file also contains metadata blocks:

and an audit block at the end:


Installation & Passive Setup

Run once after installing this skill:

# Linux / Mac / WSL
bash ~/.claude/skills/skill-index/scripts/install-hooks.sh

# Windows PowerShell
& "$env:USERPROFILE\.claude\skills\skill-index\scripts\install-hooks.sh"
# or use the PowerShell installer:
powershell -File "$env:USERPROFILE\.claude\skills\skill-index\scripts\build-index.ps1" install-hooks

This adds a SessionStart hook to ~/.claude/settings.json that runs build-index.sh update --quiet on every session start. The update operation:

  • Compares installed_plugins.json hash against the last recorded hash
  • Only re-scans if plugins changed ( ` | Register a specific skill by path |

| build-index.sh remove | Remove skill from index | | build-index.sh deprecate "" | Mark skill as deprecated | | build-index.sh verify | Validate checksums and integrity | | build-index.sh query "" | Search index (outputs matches to stdout) | | build-index.sh debug | Full diagnostic: paths scanned, errors, stats |

PowerShell equivalent: build-index.ps1 [args] — identical interface.


Scanned Locations

The scripts search for SKILL.md files in:

~/.claude/plugins/cache/*/[plugin]/[version]/skills/*/SKILL.md
~/.claude/plugins/cache/*/[plugin]/[version]/.agents/skills/*/SKILL.md
~/.claude/skills/*/SKILL.md
~/.agents/skills/*/SKILL.md

Skills missing a valid name or description in frontmatter are logged as E001 errors and skipped.


Error Codes

| Code | Cause | Remedy | |------|-------|--------| | E001 | SKILL.md missing name or description in frontmatter | Fix the skill's frontmatter | | E002 | Duplicate skill name, different paths | Check for conflicting plugins | | E003 | INDEX.md not found | Run build to create it | | E004 | SHA256 not available — CRC32 fallback active (integrity weakened) | Install sha256sum or shasum | | E005 | Plugin path in installed_plugins.json not found on disk | Reinstall the plugin |

All errors are written to ~/.claude/skills/skill-index/audit.log with timestamp and error code.


Audit Log

Every operation appends to audit.log:

2026-04-23T14:00:00Z BUILD  full_scan        total=152  duration=4.2s
2026-04-23T14:05:00Z ADD    brainstorming    v5.0.7     sha256:abc123
2026-04-23T14:06:00Z REMOVE old-skill        v1.0.0
2026-04-23T15:00:00Z UPDATE api-design       v1.9.0→v2.0.0
2026-04-23T15:01:00Z DEPRECATE legacy-skill  reason="replaced by new-skill"
2026-04-23T15:02:00Z VERIFY  integrity=valid total=153  changed=1
2026-04-23T16:00:00Z ERROR   E001            path=...   missing=description

The audit log is append-only and never modified by the scripts — safe for external monitoring (socket.dev, Snyk, CI pipelines).


Security & Auditability

  • No external dependencies — pure bash + PowerShell built-ins only
  • No network calls — filesystem operations only
  • No elevated privileges — writes only to ~/.claude/skills/skill-index/
  • Reproducible — same filesystem state always produces the same INDEX.md
  • Auditable format — plain markdown + HTML comments readable by any tool
  • socket.dev / Snyk compatible — nothing to scan (zero package dependencies)

Common Mistakes

| Mistake | Correct behavior | |---------|-----------------| | Invoking skills without reading the index first | Always Read(INDEX.md) before invoking | | Stopping after finding 1 relevant skill | Invoke ALL skills that match the task | | Using deprecated entries | Skip ~~strikethrough~~ entries | | Forgetting to run install-hooks.sh | The passive hook only works after setup | | Manually editing INDEX.md | Always use the scripts — manual edits break checksums |

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.