Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
Highly actionable with a clear, validated workflow, but the single-file structure is over-long for a skill with no external references, and the anti-premature-exit messaging is repeated across several sections.
Suggestions
Consolidate the recurring "check the clock, dispatch, don't stop early" messaging into one section (e.g. Iron Law) and have later sections reference it rather than restating it, to tighten conciseness.
Consider moving the 10-item Slop Catalog and the subagent prompt template into a reference file (e.g. references/slop-catalog.md) referenced one level deep, since the orchestrator pastes them into subagent prompts anyway.
Drop or shrink the DOT graph — the numbered step-by-step list already conveys the same dispatch loop — to remove redundancy.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly specific and earns its tokens, but the "check the clock / dispatch" mantra repeats across Your Role, Iron Law, Preventing Premature Exit, Preventing Sabotaged Runs, and Red Flags, and the DOT graph largely duplicates the step-by-step list. | 2 / 3 |
Actionability | Concrete commands (`date +%s`), exact progress-file paths and templates, a paste-ready subagent prompt, and a fixed commit message format (`anneal: <pattern> in <location>`) make the guidance copy-paste ready. | 3 / 3 |
Workflow Clarity | The Safety Protocol is an explicit sequence with validation checkpoints — run tests for baseline, make one change, re-run tests, revert on new failures, commit only when passing — a clear validate→fix→revert feedback loop. | 3 / 3 |
Progressive Disclosure | Sections are well-organized with a Quick Reference table, but the 280-line skill is monolithic; the slop catalog and subagent prompt template are inline content that could be split into one-level-deep reference files. | 2 / 3 |
Total | 10 / 12 Passed |