Content
50%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The operational core is genuinely actionable — concrete commands, endpoints, and tables — but it is buried in a file where half the content is an unrelated off-topic register, with no validation steps in its workflows and everything inlined in a single monolithic SKILL.md. Splitting the file and excising the register would raise every dimension at once.
Suggestions
Remove the ANTI-PATTERN REGISTER (and the meme-coin/high-risk-mission rules) from this skill file — it is generic agent behavior policy, not gofainder operating instructions; if it must live somewhere, put it in its own skill or reference file.
Move the full command manual, engine table, and provider table into a references/ file (e.g. references/commands.md) and keep SKILL.md as a lean overview with one-level-deep pointers.
Add validation checkpoints to the workflows, e.g. 'gofainder engines' to verify keys before search, and a curl against 127.0.0.1:8086/api/search to confirm serve is running with PORT=8086 set.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Roughly half the body (the ~100-line 'ANTI-PATTERN REGISTER' plus the golden-rule and trading-risk sections) is generic behavioral guidance unrelated to operating gofainder — much of it re-explains practices Claude already knows (write tests, handle errors, don't push prematurely). Not a 1 because the operational half (commands, tables, endpoints) is lean and table-driven; clearly below the midpoint though. | 2 / 5 |
Actionability | Concrete executable material throughout the operational sections: build commands ('go build -o gofainder ./cmd/gofainder'), exact endpoints ('GET :8086/api/search?q={query}'), systemd commands, and a per-engine key-requirements table. Not a 5 because some entries are catalog-like (module table, integration list) without runnable examples, and the PORT=8086 workaround is stated twice rather than once with a command. | 4 / 5 |
Workflow Clarity | The harvesting workflow ('research → scrape/extract → biblaigo index → godocai kb-search') and the Rules section give sequences, but there are no validation checkpoints or feedback loops anywhere (e.g., how to verify serve is up before hitting the API, or how to check engine status after configuring keys). Fits 'sequence present but checkpoints missing or implicit'. | 3 / 5 |
Progressive Disclosure | Section headers exist and no bundle files are used, but large content that clearly belongs in separate reference files is inlined: the full command manual ('Manuale d'uso approfondito'), the anti-pattern register, and the engine/provider tables all sit in SKILL.md. Fits 'some structure but could be better organized; content that should be separate is inline'; not a 2 because headers do make the file navigable. | 3 / 5 |
Total | 12 / 20 Passed |