plan-update

Use when you need to update a plan document with deep research across all compiler passes. Launches parallel subagents to analyze how a topic affects every compiler phase, then consolidates findings into the plan doc.

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./compiler/.claude/skills/plan-update/SKILL.md

Quality

Content

82%Weight 40%Scale 1-5

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

A highly actionable, well-sequenced orchestration skill: the verbatim agent prompt template and explicit phase-group partitioning make it copy-paste ready. The main gaps are minor — slight redundancy in Step 3, no failure handling for the parallel agent batch, and the pass enumeration inlined where a separate reference file could slim the main doc.

Suggestions

Add a brief validation/failure step after Step 2 (e.g., re-run or hand-cover any phase group whose agent failed or returned no usable findings) to close the workflow_clarity gap.

Tighten Step 3 by merging the prose sentence ("Update the plan document in-place to remove outdated content...") into the existing KEEP/REMOVE bullet rules to remove redundancy.

Consider moving the 55-entry pass-doc listing into a references file (e.g., references/pass-groups.md) and keeping only the phase-group names inline, reducing SKILL.md token weight.

Dimension	Reasoning	Score
Conciseness	The body is lean and operational — the 55-pass listing is data needed to partition the agent fan-out — but Step 3 carries mild redundancy ("Update the plan document in-place to remove outdated content and describe the latest understanding" overlapping the KEEP/REMOVE bullets), so it is not fully tight.	4 / 5
Actionability	Guidance is fully executable: exact context files to read, an exact agent count with phase-group partitioning, a verbatim copy-paste-ready agent prompt template with placeholders, and concrete argument examples. This matches the anchor for copy-paste-ready instructions covering the common cases.	5 / 5
Workflow Clarity	Four clearly sequenced steps with explicit consolidation rules (group by impact level, KEEP/REMOVE guidance) and a final summary checkpoint for the user. It falls short of anchor 5 because there is no validation or failure handling for the batch of 8 parallel agents (e.g., what to do if an agent fails or returns junk).	4 / 5
Progressive Disclosure	No bundle files exist, and sections are well organized with the orchestration data the main agent needs inlined. The inlined pass-document enumeration is reference-like bulk that could live in a separate file, keeping it at anchor 4 rather than a clean overview-plus-pointers structure.	4 / 5
	Total	17 / 20 Passed

Description

75%Weight 40%Scale 1-5

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

A solid, third-person description that names a distinct niche and covers both what and when with concrete actions. Its main limitation is that the when-clause mirrors the what instead of enumerating separate trigger contexts, and a few natural synonyms are absent.

Dimension	Reasoning	Score
Specificity	The description lists several concrete third-person actions — "update a plan document", "Launches parallel subagents to analyze how a topic affects every compiler phase", "consolidates findings into the plan doc" — but stops short of the comprehensive multi-capability coverage of the anchor-5 example.	4 / 5
Completeness	Both what (launch parallel subagents, analyze, consolidate) and when ("Use when you need to update a plan document with deep research across all compiler passes") are present, but the when-clause largely restates the what rather than naming distinct triggering situations, so it is not fully explicit.	4 / 5
Trigger Term Quality	Natural phrases like "update a plan document", "deep research", and "compiler passes" match what a user would plausibly say, though common variations ("plan doc", "pass impact", "ripple effects") are missing.	4 / 5
Distinctiveness Conflict Risk	The compiler-pass/plan-document niche is distinct with minimal overlap risk, but the "deep research" phrasing could attract generic research requests that belong to a general research skill.	4 / 5
	Total	16 / 20 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: facebook/react
Path: compiler/.claude/skills/plan-update/SKILL.md
Commit: 7c6ac13

Reviewed: 3 days 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.