Content
86%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.
A well-structured overview skill: opinionated domain guidance, hard quality gates, and clean routing to a complete set of one-level-deep reference files, all of which exist. The main improvement opportunities are consolidating the duplicated reference routing between 'The PR Mix' and 'Common Workflows' and making the fix-and-recheck loop after the pitch quality bar explicit.
Suggestions
Consolidate the duplicated reference pointers: 'The PR Mix' and 'Common Workflows' both route to newsjacking.md, journalist-pitching.md, story-angles.md, press-platforms.md, and podcast-guest-prep.md — either merge the routing into one section or make the PR Mix table link directly and drop the repeated 'For the X — see Y' lines.
Close the feedback loop on the pitch quality bar: after 'If any answer is no, don't send', add the recovery step (e.g., 'Revise the pitch against the failed check and re-run the full checklist before sending') so validation cycles back instead of just halting.
Tighten the first-60-minutes crisis checklist in the body into a numbered sequence (or point to the exact checklist section in crisis-communications.md) so the 'When Something Goes Wrong' workflow reads as an ordered runbook rather than a prose summary.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes competence — it teaches PR strategy judgments ('AVE ... is a vanity metric PR firms invented', 'Speed beats polish on reactive PR') rather than generic concepts, and uses tables and checklists instead of prose. However, 'The PR Mix' section and 'Common Workflows' section both route to the same five reference files (newsjacking.md, journalist-pitching.md, story-angles.md, press-platforms.md, podcast-guest-prep.md are each cited twice), a minor redundancy that could be consolidated. This sits between anchors 4 and 5, closer to 4: efficient with minor trimmable instances, not 'every token earns its place.' | 4 / 5 |
Actionability | Guidance is directly executable for an instruction-only skill: the 7-item pitch quality bar with hard thresholds ('Is the pitch under 150 words?', 'Check their last 5 articles.', banned-word list, 'If any answer is no, don't send'), the 9-item press page checklist, the HARO rule ('keep it under 200 words'), the 24-hour response commitment, and the measurement table. Specific procedures cover the common cases for each workflow, matching anchor 5; there are no gaps that force guessing. | 5 / 5 |
Workflow Clarity | Each workflow is clearly sequenced ('run the scoring rubric, draft 2–3 angles, pick the best, draft the pitch'; crisis: 'assign a severity tier, then run the first-60-minutes checklist: confirm facts, designate one spokesperson, publish a holding statement, set the next update time') and includes validation checkpoints ('If any answer is no, don't send', 'loop in legal counsel before admitting fault or publishing'). However, the error-recovery half of the loop is implicit — the quality bar says don't send but never says revise-and-recheck — which keeps it at anchor 4 ('clear sequence with most checkpoints present; minor validation gaps') rather than 5 ('feedback loops for error recovery'). | 4 / 5 |
Progressive Disclosure | The body is a genuine overview: all 8 bundle files in references/ are cited with a one-line purpose ('For the reactive newsjacking workflow — see references/newsjacking.md'), every cited path exists on disk, and links are one level deep with only lateral sibling cross-links (e.g., crisis-communications.md ↔ statement-templates.md), which aid navigation rather than bury content. This matches anchor 5 ('clear overview with well-signaled one-level-deep references; easy navigation'). | 5 / 5 |
Total | 18 / 20 Passed |