Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The content is highly actionable with a clearly sequenced, well-validated multi-step workflow, but it is a long monolithic file with light redundancy between its safety sections and inline content that could be factored into references.
Suggestions
Consolidate the overlapping untrusted-input guidance between 'Trust Boundaries and Command Safety' and 'Safety Rules' into a single section to remove redundancy and save tokens.
Move the multi-line GraphQL query blocks and the bot-login allowlist into reference files (e.g. references/graphql-queries.md, references/bot-logins.md) with one-level-deep links from the body to improve progressive disclosure.
Tighten Step 5's filtering narrative into a compact checklist so the retention rules read as concisely as the rest of the workflow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and prescriptive with no concept-explanation fluff, but it repeats the untrusted-input guidance in both 'Trust Boundaries and Command Safety' and the closing 'Safety Rules', introducing minor redundancy that keeps it just below the leanest anchor. | 2.5 / 3 |
Actionability | Concrete, executable `gh api`, `git`, and GraphQL commands with quoted variables and `--` pathspecs appear throughout, making the guidance copy-paste ready. | 3 / 3 |
Workflow Clarity | Twelve sequenced steps include explicit validation checkpoints (git-state checks in Step 2.5, remote merge-base verification in Step 10, re-checking staged files before commit) with feedback loops for the destructive push/resolve operations. | 3 / 3 |
Progressive Disclosure | It is a single ~375-line SKILL.md with well-organized sections but no bundle files; sizable inline blocks such as the GraphQL queries and the bot-login list could be split into references, so it sits at the 'some structure but inline-heavy' anchor. | 2 / 3 |
Total | 10.5 / 12 Passed |