BB repo-explainer
Decode a GitHub repository **or a local code directory** into a non-developer-friendly report, delivered as BOTH a polished HTML page AND a Markdown document. Given either a GitHub repo URL or a local directory path, this skill prepares the source (shallow-clone for URLs, in-place for local paths), scans the ACTUAL source code (not just README), and produces two synchronized reports containing: project overview, tech stack table, architecture diagram, key module table, main user-flow sequence diagram, API/CLI surface table, third-party dependency table, and a "what this project actually does in plain language" section. The HTML is themed and print-ready; the Markdown embeds native Mermaid blocks and pastes cleanly into Feishu / Notion / GitHub. Every conclusion is backed by `path:line` references — for GitHub URLs (and local clones with a github remote), references render as clickable permalinks anchored to the commit SHA; for non-git local dirs they render as monospace text. TRIGGER WHEN the user gives a `github.com/<owner>/<repo>` URL **OR** a local directory path (absolute / relative / `~`-prefixed) AND asks any of: "解读这个项目"、"分析这个仓库"、"分析这个目录"、"讲讲这个 GitHub 项目"、 "帮我看下这个 repo / 这个文件夹做什么"、"explain this repo"、 "analyze this github project"、"analyze this folder"、 "what does this repository do"、"understand this codebase"、 "画一下这个项目的架构"。 DO NOT TRIGGER for: private/internal GitLab URLs, single-file gists, or when the user only wants to run/install the project (not understand it). Local-directory mode IS in-scope — including monorepos (use `--subpath`) and not-yet-pushed work-in-progress checkouts.
Decode a GitHub repository or a local code directory into a non-developer-friendly report, delivered as BOTH a polished HTML page AND a Markdown document.
As a process B 71/100 · Nearly there — weak spots: result and completion, running it twice, progress reporting
How to improve
- Shorten the description to 1024 characters.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 11. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1621 chars, limit 1024 - note
description-budgetdescription takes 1621 of the ~15000-char shared budget for all skills
Process rating: all ten parameters 71/100
- 0Result and completion. Does not say what the result is
- 0Progress reporting. Says nothing while it works
- 30Running it twice. 7 mutating operations with no state check
- 60Tools and files. Uses tools (bash, web, git, node) that frontmatter does not declare
- 70Inputs and preconditions. Inputs and preconditions are listed
- 100Steps. 55 steps
- 100When it triggers. States when to use and when not to
- 100Failures and branches. 1 branches, has a failure section
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3240 tokens
- low The response is described with custom markup (12 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 1620: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -5TODO / placeholder text left in the skill
- -4Absolute local paths (C:\Users, /home/…): not portable
- +1No license
- +2Single-language instructions
- +5Description quotes 11 example trigger phrases
- +4Structure: 11 headings
- +3Step-by-step instructions: 55 items
- +4Has examples (8 code blocks)
- +3All 5 scripts are documented
Quality base 70; lint remarks subtract, signals add up to 100. Result: 52.