Skip to main content
hc
Submit
skillsSource-backedReview first Safety ✓ Privacy ✓

Agent Skills Retrieval Source Verification Capability Pack Skill

Expert agent skills retrieval source verification capability pack for validating canonical URLs, documentation reachability, repo provenance, and duplicate source claims before publishing or submitting Claude Code skills.

by kiannidev·added 2026-06-14·
Claude CodeCodexWindsurfGeminiCursorCLI
HarnessClaude CodeCodexWindsurfGeminiCursorCLI
Level:expertType:capability-packVerified:validated
Review first — review before installing

Open the source and read safety notes before installing.

Safety notes

  • Retrieval sources must point to primary official documentation or verifiable repositories, not affiliate landing pages or unreviewed mirrors.
  • Parked domains, empty documentation pages, and redirect chains to unrelated products fail source verification.
  • Listing many URLs without explaining which is canonical creates reviewer ambiguity and increases duplicate risk.
  • This skill verifies sources; it must not approve a submission whose claims cannot be substantiated by fetched evidence.

Privacy notes

  • Source verification logs may include internal staging URLs, private repo names, or authentication-gated docs if reviewers paste raw fetch output.
  • Some documentation pages embed organization-specific examples that should be redacted before public issue comments.
  • Public verification summaries should list reachability status and canonical URL choice, not full HTTP response bodies.

Prerequisites

  • Draft skill frontmatter with `retrievalSources`, `repoUrl`, and `documentationUrl` fields populated.
  • Network access or cached evidence to verify HTTP reachability of listed source URLs.
  • The skill's claimed scope, slug, and category for duplicate comparison against existing entries.
  • Contributor or reviewer context for whether docs, repo, or package is the canonical source of truth.

Schema details

Install type
package
Reading time
8 min
Difficulty score
78
Troubleshooting
Yes
Breaking changes
No
Source repository stats
Scope
Source repo
Package metadata
Skill and platform metadata
Skill type
capability-pack
Skill level
expert
Verification
validated
Verified at
2026-06-14
Retrieval sources
https://code.claude.com/docs/en/skillshttps://code.claude.com/docs/en/features-overviewhttps://github.com/anthropics/claude-codehttps://developers.google.com/search/docs/fundamentals/creating-helpful-content
Tested platforms
ClaudeClaude CodeCodexCursorWindsurfGeneric AGENTS
PlatformSupportInstall path
claude-codeNative.claude/skills/<skill-name>/SKILL.md
codexNative.agents/skills/<skill-name>/SKILL.md
windsurfNative.windsurf/skills/<skill-name>/SKILL.md
geminiNative.gemini/skills/<skill-name>/SKILL.md or .agents/skills/<skill-name>/SKILL.md
cursorAdapter.cursor/rules/<skill-name>.mdc
cliManualAGENTS.md or tool-specific context file
Full copyable content
# Trigger
"Apply the agent skills retrieval source verification capability pack."

# Required output
1) Canonical source URL inventory and reachability results
2) Provenance and duplicate-source assessment
3) Documentation versus repo truth alignment review
4) Retrieval source quality and freshness notes
5) Privacy-safe verification summary

About this resource

Knowledge Freshness

This capability pack is grounded in Claude Code skills and features overview documentation verified on 2026-06-14. Source verification requirements and submission gate behavior can evolve; prefer live official docs and reachable primary sources over cached URL assumptions.

Retrieval Sources

Source Verification Notes

Verified against official Claude Code skills documentation and the public Anthropic claude-code repository on 2026-06-14:

  • Claude Code skills are defined in SKILL.md files with frontmatter that can include descriptions, allowed tools, and supporting references.
  • Skill descriptions load into Claude Code context at session start when model invocation is enabled, so inaccurate source claims affect every session.
  • HeyClaude content submissions require verifiable canonical sources; thin, parked, or unreachable documentation URLs trigger manual review or closure.
  • Using the public anthropics/claude-code repository as documentationUrl provides a stable, HTTP-verifiable canonical source when topic docs live on separate documentation hosts listed in retrievalSources.
  • Google Search Central helpful-content guidance supports evaluating whether skill copy is useful, specific, and source-backed rather than promotional.

Scope Note

This is not a web crawling or legal provenance service. Use it as a reusable verification workflow for authors and reviewers validating agent skill retrieval sources before HeyClaude submission or internal skill publication.

Core Workflow

  1. Inventory all source fields: documentationUrl, repoUrl, downloadUrl, and every entry in retrievalSources.
  2. Select one canonical source of truth and ensure other URLs support rather than contradict it.
  3. Verify reachability: HTTP 200 for public docs and repos, note redirects, parked pages, auth walls, and final resolved URLs.
  4. Compare skill claims to fetched evidence: feature names, commands, flags, and workflows must appear in official docs or repo README.
  5. Check duplicate history: search existing skills by slug, title, repo URL, docs domain, and overlapping purpose.
  6. Assess freshness: record verification date in Knowledge Freshness and flag time-sensitive features such as beta flags or enterprise-only scope.
  7. Review privacy and safety notes against source truth; do not invent runtime guarantees absent from official documentation.
  8. Document unresolved gaps explicitly rather than substituting forum posts or unverified social threads as primary sources.
  9. Produce a verification summary suitable for PR body or gate review.

Capability Scope

  • Canonical source URL selection and reachability checks.
  • Provenance and duplicate-source assessment.
  • Documentation versus repository truth alignment.
  • Retrieval source quality and freshness review.
  • Submission-ready verification summary.
  • Privacy-safe reporting of fetch results.

Compatibility

Native

  • Claude Code / Claude: use as an Agent Skill when authoring SKILL.md files, preparing HeyClaude submissions, or reviewing contributor PRs.

Manual Adaptation

  • Codex, Cursor, Windsurf, and Generic AGENTS workflows: use the workflow as a deterministic source-verification checklist in content review runbooks.

Required Inputs

  • Draft skill frontmatter and body with claimed capabilities and source URLs.
  • Reachability evidence for each listed URL (status code, final URL, date).
  • Duplicate search results across content/skills/ and open PRs.
  • Category, slug, and intended audience for the skill under review.

Production Rules

  • Prefer one canonical verifiable URL in documentationUrl; list topic docs in retrievalSources.
  • Do not use affiliate, referral, or paid listing pages as primary sources.
  • Do not approve skills whose core claims lack official documentation support.
  • Record verification date and concrete fetched facts in Source Verification Notes.
  • Redact auth tokens, internal URLs, and customer examples from public summaries.
  • Close or revise submissions with parked docs rather than retrying identical URLs.
  • Separate helpful-content quality review from reachability checks; both matter.

Review Matrix

Signal Pass Fail Action
Canonical URL Official repo or live docs Parked or empty page Change canonical source
Reachability HTTP 200 public 404 or auth wall Fix URL or scope claim
Claim alignment Supported by source Unsupported feature Revise skill body
Duplicates Distinct slug/purpose Same repo/title/workflow Close or differentiate
Freshness Dated verification Stale beta claim Update Knowledge Freshness
Promo tone Practical workflow Affiliate copy Rewrite or close

Output Contract

  1. Canonical source URL inventory and reachability results.
  2. Provenance and duplicate-source assessment.
  3. Documentation versus repo truth alignment review.
  4. Retrieval source quality and freshness notes.
  5. Pass/fail recommendation with required fixes.
  6. Privacy-safe verification summary for PR or gate review.

Duplicate Check

Checked content/skills, content/guides, generated catalog text, and open pull requests for agent skills retrieval source verification, canonical URL review, and HeyClaude source verification workflows. Official submission docs describe source requirements, but no skills entry provides a reusable retrieval source verification capability pack with review matrix and output contract.

Editorial Disclosure

Submitted as an independent source-backed HeyClaude content entry by kiannidev. It is based on public Claude Code documentation, the public Anthropic claude-code repository, and Google Search Central helpful-content guidance. No paid placement, referral link, affiliate link, or vendor sponsorship is used.

#agent-skills#retrieval-sources#source-verification#provenance#capability-pack

Source citations

Add this badge to your README

Show that Agent Skills Retrieval Source Verification Capability Pack Skill is listed on HeyClaude. Paste this Markdown into your README — it renders the badge and links back to this page.

Listed on HeyClaude
[![Listed on HeyClaude](https://heyclau.de/badge/skills/agent-skills-retrieval-source-verification-capability-pack.svg)](https://heyclau.de/entry/skills/agent-skills-retrieval-source-verification-capability-pack)

Signals

Loading live community signals…

More like this, weekly

A short, calm digest of reviewed Claude resources. Unsubscribe any time.