Content
85%Weight 40%Scale 1-3Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, actionable instruction skill: an explicit sequenced workflow with validation gates, concrete paths and connector names, and clean one-level-deep references to a real bundle. The main weakness is verbatim repetition of the registry-propose protocol and red-line motifs across sections, which costs conciseness.
Suggestions
Consolidate the 'authorized operation: propose request to registry-events.py' explanation into one place (e.g. the Skill Contract Writes line) and reference it from steps 3, 8, and Save Results rather than re-stating it each time.
State the M1 veto / red-line once in step 7 and link from the intro and scope guard instead of re-describing it in three sections.
Move the 'verify current rules via WebFetch' rationale (currently repeated in Data Sources, step 2, and the channel-matrix reference) into a single sentence to trim repeated tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and assumes Claude's competence (no padding explaining what a subreddit or Product Hunt is), but several motifs repeat verbatim 2–3 times — the 'authorized operation: propose request to registry-events.py' protocol (steps 3, 8, Save Results), the M1 veto / red-line (intro, scope guard, step 7, references), and 'never writes the dossier directly'. Mostly efficient but could be tightened; not the lean 'every token earns its place' level. | 2 / 3 |
Actionability | Concrete and executable for an instruction skill: exact paths ('memory/events/launches.ndjson', 'memory/launch/community-launch-runner/'), named connectors ('scripts/connectors/hn.py', 'producthunt.py', 'gdelt.py'), explicit title format ('Show HN: <what it is, stated plainly>'), and field specs cited to official docs marked verify-current. Per scoring_notes, absence of code is not penalized when guidance is this actionable. | 3 / 3 |
Workflow Clarity | An 8-step numbered process with explicit validation gates: verify each community's current rules via WebFetch before committing (step 2), drop channels whose rules bar self-promotion, ask mods when ambiguous, the red-line check (step 7), and a 'Done when' completion criterion. Clear sequence with checkpoints and feedback loops. | 3 / 3 |
Progressive Disclosure | Well-organized sections (Quick Start, Skill Contract, Data Sources, Instructions, Save Results, Reference Materials, Next Best Skill) with one-level-deep, clearly signaled references including the real bundle file references/channel-matrix.md (confirmed present), plus ramp-benchmark.md, CONNECTORS.md, SECURITY.md, and skill-contract.md. No nested-reference chains. | 3 / 3 |
Total | 11 / 12 Passed |