node-check

Verify one coder node against the assigned execution contract and return pass, rework, blocked, or non-converged evidence.

Quality

68%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./docs/plantree/plans/agentic-loop-workflow/drafts/agentroles.code_reviewer/skills/node-check/SKILL.md

Quality

Content

86%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.

The body is an exemplary lean, role-style instruction set: a numbered verification workflow with an unambiguous machine-readable verdict contract and firm boundaries, with no wasted tokens. Its only real weakness is that a few audit directives tell the reviewer what to look for without showing how.

Dimension	Reasoning	Score
Conciseness	The body is a lean 7-step workflow plus a boundaries list with zero padding: it never explains concepts Claude already knows, and compressed directives like "Canonical tree digest is dispatcher/controller-only route evidence" assume the model's competence and leave out all filler. Every token carries instruction weight, matching 'lean and efficient; every token earns its place'.	5 / 5
Actionability	The output contract is copy-paste precise ("first non-empty line: `status: pass`, `status: rework_required`, `status: blocked`, or `status: non_converged`") and several steps name exact checks (allowed paths, undeclared dependencies, read-only verification). It falls short of 5 because steps like "Audit hidden fallback, degradation, scope shrinkage, and missing evidence" give direction without any method or example of what such evidence looks like — minor gaps, fitting 'mostly executable guidance with minor gaps' rather than the incomplete anchor 3.	4 / 5
Workflow Clarity	The seven steps form a coherent sequence with explicit decision checkpoints ("Missing or mismatched visible identity is `blocked`", "otherwise report the missing proof", "Put explanatory evidence after that line"). It does not reach 5 because there is no feedback loop for error recovery and some checkpoints (how to detect a scope violation) are named but not elaborated — 'clear sequence with most checkpoints present; minor validation gaps'.	4 / 5
Progressive Disclosure	The skill is under 50 lines with two well-organized sections (Workflow, Boundaries) and no content that warrants external reference files — no references/scripts/assets exist in the bundle. Per the rubric's simple-skill guidance, well-organized sections alone merit 5.	5 / 5
	Total	18 / 20 Passed

Description

51%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.

The description communicates a concrete, well-bounded capability with a crisp set of verdict outcomes, but it reads as internal pipeline documentation rather than a skill description: it is jargon-heavy, lacks any natural trigger terms, and omits the 'when to use' half entirely.

Suggestions

Add an explicit trigger clause, e.g. 'Use when a Worker supplies a node review packet for verification, or when checking a completed node against its execution contract.'

Replace or gloss the internal jargon ('coder node', 'non-converged evidence') with terms a user would naturally say, such as 'review completed work', 'accept or reject', 'verify against contract'.

Name one or two of the concrete checks performed (e.g. path/scope validation, read-only verification of test evidence) so the 'what' covers the means, not just the outcomes.

Dimension	Reasoning	Score
Specificity	"Verify one coder node against the assigned execution contract and return pass, rework, blocked, or non-converged evidence" lists a concrete action plus a fully enumerated set of outcome states (pass, rework, blocked, non-converged), matching 'several specific actions; minor gaps'. It is not a 5 because it stays at the level of outcomes without covering what the verification actually involves.	4 / 5
Completeness	The 'what' is clear (verify a node against its contract and return one of four verdicts), but there is no 'when' clause at all — no 'Use when...' or equivalent trigger guidance — which caps completeness at 3 per the judging guidelines. It is not a 2 because the 'what' half is concrete, and not a 4 because the 'when' is entirely absent rather than just weak.	3 / 5
Trigger Term Quality	The description is built almost entirely from pipeline jargon — "coder node", "execution contract", "non-converged evidence" — with no natural phrases a user would say when they need this skill. Only the generic word "verify" is user-natural, fitting 'one or two generic keywords; missing the natural phrases users say' rather than the jargon-only anchor 1.	2 / 5
Distinctiveness Conflict Risk	The niche framing (bounded node review packets, machine verdicts for a controller pipeline) is clearly distinguishable from generic code-review or QA skills with minimal overlap risk. It falls short of 5 only because the absence of trigger phrases means 'verify' could loosely attract unrelated verification requests.	4 / 5
	Total	13 / 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: SeemSeam/claude_codex_bridge
Path: docs/plantree/plans/agentic-loop-workflow/drafts/agentroles.code_reviewer/skills/node-check/SKILL.md
Commit: 18bf0af

Reviewed: 7 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.