Audit the Skills in a Repository

Problem/Feature Description

A team-mate has dropped off a small repository and asked you to give them a one-page summary of every skill it contains: how many there are, how stale each one looks, whether any of them duplicate each other, and how their content quality stacks up. They specifically want a single self-contained HTML report they can share in Slack, and the underlying JSON files so they can grep through the raw data afterwards.

The repository lives at ./resources/myrepo in your working directory. It is a real git checkout (the setup script you were handed has already initialised it and made one commit so that the history is non-empty).

Output Specification

Run the full skill-insights pipeline against ./resources/myrepo and produce, in the repo's .skill-insights/ directory, all of:

• discovery.json — the canonical inventory
• staleness.json — per-skill staleness scores
• quality.json — per-skill quality scores (or a graceful failure record if the Tessl CLI is unavailable in this sandbox)
• duplicates.json — duplicate clusters and overlap pairs
• report.html — the rendered self-contained report

Once the pipeline has run, write a short summary to ./pipeline-log.md (in the working directory, NOT inside the repo) that records:

1. The order in which you ran the phases (which were sequential, which were parallel)
2. The final skill count and repo count from discovery.json
3. Whether any phase degraded gracefully and why