Tier-2 Review: skill-seekers (Cluster: document-generation)

What We Attempted

We evaluated skill-seekers, a skill sourced from SkillsMP that claims to automatically convert documentation websites, GitHub repositories, and PDFs into Claude AI skills. The skill presents itself as a high-level automation tool for transforming existing documentation into actionable Claude skills, with a composite score of 4.5/5.0 across all five dimensions (trigger clarity, output specificity, scope precision, self-containment, and reusability).

Our test harness attempted three standard validation paths: a full install sequence to verify the skill can be set up in a fresh Python 3.10+ environment, a minimal-roundtrip test to confirm a basic file transformation works end-to-end, and an edge-corner test to assess resilience with non-standard inputs (e.g., scanned PDFs lacking machine-readable text).

What Failed, and What We Observed

All three tests failed cleanly. No tests passed, no partial results were achieved. The failures are not due to environmental flakiness but stem from structural gaps in the skill's documentation and design.

1. Install Failure: The SKILL.md file contains no installation instructions whatsoever. There is no requirements.txt, no setup.py, no pip command, and no mention of dependencies beyond a Python 3.10+ requirement implied by the source. A user cannot set up this skill without reverse-engineering the codebase or guessing dependencies. This is a critical omission for any production-ready skill.

2. Minimal-Roundtrip Failure: The skill's stated purpose is converting documentation websites, GitHub repos, and PDFs into Claude skills. However, no minimal transformation command is defined—there is no example of "take file X, run command Y, get output Z." The SKILL.md describes the what but not the how. Without a clear, repeatable transformation pathway, automated testing cannot validate that the skill does anything useful.

3. Edge-Corner Failure: The skill claims to handle PDFs, but it does not mention OCR capabilities. A scanned PDF (image-based, no embedded text) would fail silently or produce garbage output. The skill also only targets documentation websites and GitHub repos—there is no fallback for non-standard formats, malformed inputs, or edge cases like password-protected PDFs or offline repositories.

Rating: Theoretically Sound, Practically Unverified

The composite score of 4.5/5.0 is therefore theoretical until a physical re-run resolves these failures. The skill's concept—automating skill creation from existing docs—is genuinely valuable, and the high scores across all five dimensions suggest the authors thought deeply about clarity, specificity, scope, self-containment, and reusability. However, the complete absence of install instructions and transformation commands means the skill cannot be deployed or tested as-is.

Is the Skill Still Valuable in Principle?

Yes, the idea remains strong. Automating the conversion of documentation websites, GitHub repos, and PDFs into Claude skills would save significant manual effort for developers and technical writers. The skill's scope is well-defined (documentation sources only), and the output specificity (Claude skills) is clear. If the authors add a proper install path, a minimal example command, and note OCR limitations, this skill could justify its high score. Until then, it is a promising but non-functional artifact.

Recommendation: The maintainer should add a requirements.txt, document a single convert command with a working example, and explicitly state that scanned PDFs are not supported. Once those three gaps are closed, a re-run would likely yield passing tests.

Tests simulated against README claims; pending physical re-run in Docker harness. Ran 2026-06-26.

Overall: broken. 0 tests passed, 0 partial, 3 failed; key blocker: SKILL.md lacks install instructions, transformation commands, and edge-case handling details.

Inferred dependencies: python>=3.10.

Test	Status	Notes
install	fail	No install command documented in SKILL.md; missing requirements.txt or setup.py. Assumes Python environment but no dependency listing.
minimal-roundtrip	fail	SKILL.md describes converting websites and repos, not a file roundtrip. No minimal transformation defined.
edge-corner	fail	No OCR capability mentioned; scanned PDF would fail. Skill only handles documentation websites, GitHub repos, and PDFs without OCR.