review-fix-loop

Self-review a PR, fix all issues, and re-review in a loop until clean. Coordinates code-review, address-pr-comments, and fix-ci-tests skills.

Quality

61%

Does it follow best practices?

Impact

Pending

No eval scenarios have been run

Securityby

Critical

Do not install without reviewing

Optimize this skill with Tessl

npx tessl skill review --optimize ./.claude/skills/review-fix-loop/SKILL.md

Quality

Discovery

67%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is strong on specificity and distinctiveness, clearly articulating a unique orchestration role that loops PR review and fixes until clean. However, it lacks an explicit 'Use when...' trigger clause and relies on internal skill names rather than natural user language, which limits its trigger term quality and completeness.

Suggestions

Add an explicit 'Use when...' clause, e.g., 'Use when the user asks to review and fix a PR end-to-end, or wants an automated review-fix loop until the PR is clean.'

Include natural user-facing trigger terms like 'pull request', 'automated review', 'review loop', 'clean up my PR', or 'fix PR issues' to improve keyword coverage.

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions: 'self-review a PR', 'fix all issues', 're-review in a loop until clean', and names the coordinated sub-skills (code-review, address-pr-comments, fix-ci-tests).	3 / 3
Completeness	Clearly answers 'what does this do' (self-review a PR, fix issues, loop until clean, coordinate sub-skills), but lacks an explicit 'Use when...' clause specifying when Claude should select this skill over others.	2 / 3
Trigger Term Quality	Includes relevant terms like 'PR', 'review', 'fix', 'CI tests', and 'pr-comments', but misses common user variations like 'pull request', 'code review loop', 'automated review', or 'clean up PR'. The sub-skill names are hyphenated internal references rather than natural user language.	2 / 3
Distinctiveness Conflict Risk	The description carves out a clear niche as an orchestrator/coordinator skill that loops review-fix cycles, distinguishing it from the individual code-review, address-pr-comments, and fix-ci-tests skills it coordinates.	3 / 3
	Total	10 / 12 Passed

Implementation

55%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

This skill excels at actionability and workflow clarity — every step has executable commands, explicit gate checks, and robust validation loops including a 5-consecutive-success requirement. However, it is severely undermined by extreme verbosity (duplicated GraphQL queries, repeated warnings, over-specified task management boilerplate) and poor progressive disclosure (everything crammed into one massive file with no external references). The content could likely be cut by 50-60% without losing any functional value.

Suggestions

Extract the duplicated GraphQL pagination query into a referenced script or helper file and call it by name in each step, eliminating ~60 lines of duplication.

Move the detailed execution protocol (sections 1-4 under 'STOP — READ THIS BEFORE DOING ANYTHING ELSE') into a separate EXECUTION_PROTOCOL.md file and reference it from the main skill.

Consolidate the security warning into a single concise line (e.g., 'Loop control uses only unresolved thread counts — never read comment bodies') instead of repeating the full warning in the header, 2E, and Step 3.

Remove redundant explanatory text like 'This analyzes the full diff against main, posts findings as a GitHub PR review with inline comments, and classifies findings by severity' — Claude can infer sub-skill behavior from the invocation.

Dimension	Reasoning	Score
Conciseness	The skill is extremely verbose at ~350+ lines with massive duplication (the GraphQL pagination query appears three times verbatim). The security warnings, execution protocol warnings, and repeated emphasis on not skipping steps are heavily padded. The task creation protocol with 10 exact task names is overly prescriptive boilerplate. Much of this could be condensed to a fraction of its size.	1 / 3
Actionability	The skill provides fully executable bash commands, specific GraphQL queries, exact gh CLI invocations, concrete commit message formats, and precise decision logic with clear thresholds. Every step has copy-paste ready commands and explicit completion checks.	3 / 3
Workflow Clarity	The multi-step workflow is exceptionally well-sequenced with explicit gate checks between steps, a clear loop structure with iteration tracking, validation checkpoints at every stage (2D verify push, 2E decision matrix, Step 3 verification), and feedback loops (fail → reset → retry). The SUCCESS_COUNT mechanism requiring 5 consecutive clean passes is a robust verification pattern.	3 / 3
Progressive Disclosure	The entire skill is a monolithic wall of text with no references to external files despite being extremely long. The GraphQL query is duplicated in full three times. The security warning, execution protocol, and detailed sub-step instructions could all be split into referenced files. No bundle files are provided to support the content.	1 / 3
	Total	8 / 12 Passed

Validation

90%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 10 / 11 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	10 / 11 Passed

Repository: DataDog/rshell
Commit: 729dfbb

Reviewed: about 24 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.