Content
81%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 content is a strongly actionable, well-sequenced workflow with genuine validation and error-recovery checkpoints, and a complete output template. Its weaknesses are mild padding in the generic insufficient-input preamble and repeated minimal-workflow conditionals, plus a monolithic single-file structure where the output template could be split into a reference file.
Suggestions
Trim the "Insufficient input" preamble to its operative rules; drop the off-domain asset-audit/performance-profile examples that do not concern retrospectives.
Consolidate the `workflow: minimal` / `qa.level: minimal` conditionals into one place instead of restating them in Phase 2 and Phase 6.
Move the Phase 4 markdown output template into a references/ file (e.g., references/retro-template.md) and link to it, keeping SKILL.md as the workflow overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly operational — exact paths, glob patterns, prompts, commands, and a full output template — but the ~40-line "Insufficient input" preamble pads with generic, off-domain examples ("an asset audit returning COMPLIANT", "performance profile reporting >99% headroom") irrelevant to retrospectives, and the `workflow: minimal` / `qa.level: minimal` handling repeats across Phases 2 and 6. It is tighter than anchor-2 verbosity but has several sections that could be trimmed, matching anchor 3. | 3 / 5 |
Actionability | Guidance is fully executable: a concrete git log command (with the Git-Bash/PowerShell pitfall spelled out), exact glob patterns for existing retro detection, file paths for every data source, verbatim AskUserQuestion prompts with lettered options, and a complete copy-paste markdown output template covering the common cases. Nothing is pseudocode or hand-waved. | 5 / 5 |
Workflow Clarity | Six phases are clearly sequenced with explicit validation checkpoints: a FOUND/ABSENT input protocol before any report, an existing-retrospective check with user-chosen recovery (update vs archive), data-existence gates with BLOCKED verdicts, and a permission ask gating every write (including the archive rename). Error paths and NOT ASSESSED — NO DATA states are explicitly defined, matching the anchor-5 pattern of explicit validation plus feedback loops. | 5 / 5 |
Progressive Disclosure | The body is a single well-organized file with clear phase-level headers and no buried or nested references (none exist), so navigation is easy. It earns 4 rather than 5 because at ~280 lines the large Phase 4 output template (~90 lines) is a natural candidate for a one-level-deep references/ file (e.g., references/retro-template.md), which would slim SKILL.md to its workflow; there is no well-signaled reference structure to reward with a 5. | 4 / 5 |
Total | 17 / 20 Passed |