agent-harness-diagnose

Diagnostica la causa raiz de un arnes que deriva: reporte de drift docs-vs-realidad (claims sin entradas, entradas huerfanas, evidencia rota, comandos no verificables, placeholders residuales) via reconcile determinista. Usar al pedir 'diagnosticar arnes', 'harness drift', 'progress desactualizado', 'el arnes miente'. NO para reparar (agent-harness-repair) ni para puntuar (agent-harness-audit).

Quality

85%

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

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 concise, well-structured diagnostic skill body with an executable entry command and clear sequencing plus a validation signal. Gaps are an un-templated output format, no error-recovery loop, and a Packet block that overstates the bundled layer folders.

Suggestions

Add a short template or example of the expected diagnostic output (the 'cadena causal' and per-bucket 'siguiente paso') so the deliverable is copy-paste concrete, not just described.

Surface scripts/check.sh from the body (e.g., a 'Self-check' line in Procedimiento or Packet) so the bundled verification script is discoverable.

Reconcile the Packet block with the actual bundle: either add the missing layer folders (knowledge/, prompts/, examples/, agents/) or trim the list to the folders that exist (references/, scripts/, assets/) to avoid overstating structure.

Dimension	Reasoning	Score
Conciseness	Lean and dense throughout ('Encuentra POR QUE un arnes dejo de reflejar la realidad'), assumes Claude's competence, explains no basic concepts, and uses tagged contract lines ([CÓDIGO], [EXPLICIT]) instead of prose padding.	5 / 5
Actionability	Provides one copy-paste-ready command ('reconcile_harness.py <dir> --json') and concrete classification steps, but only a single command is given and the expected diagnostic output format ('cadena causal') is not templated.	4 / 5
Workflow Clarity	A clear four-step Procedimiento with the reconcile exit code as an explicit validation signal ('exit 1 del reconcile reflejado como drift confirmado'); read-only so the destructive/batch cap does not apply, but there is no explicit error-recovery feedback loop.	4 / 5
Progressive Disclosure	Well-signaled one-level-deep reference to a real bundle file ('references/drift-report.md') and clear section structure, but scripts/check.sh exists yet is unreferenced and the Packet block lists layer folders (knowledge/, prompts/, examples/, agents/) absent from the actual bundle.	4 / 5
	Total	17 / 20 Passed

Description

87%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 strong, specific description that concretely states what it does, when to use it with natural trigger phrases, and how it differs from sibling skills. Minor room to broaden trigger-term synonyms and split the single macro-action into more granular capabilities.

Dimension	Reasoning	Score
Specificity	Names the domain and enumerates six concrete drift buckets ('claims sin entradas, entradas huerfanas, evidencia rota, comandos no verificables, placeholders residuales') plus the mechanism 'via reconcile determinista', but the core action is a single diagnose-and-report macro-action rather than multiple distinct actions.	4 / 5
Completeness	Clearly and explicitly answers both 'what' (diagnose root cause of a drifting harness via a six-bucket drift report) and 'when' (concrete trigger phrases), supplemented by explicit negative boundaries.	5 / 5
Trigger Term Quality	Provides four natural user-facing trigger phrases ('diagnosticar arnes', 'harness drift', 'progress desactualizado', 'el arnes miente') with explicit 'Usar al pedir' guidance, though some synonym and English/Spanish variant coverage is missing.	4 / 5
Distinctiveness Conflict Risk	Clear niche (harness drift diagnosis) with distinct triggers and explicit disambiguation from sibling skills ('NO para reparar (agent-harness-repair) ni para puntuar (agent-harness-audit)'), plus routes-to/composes-with metadata minimizing conflict risk.	5 / 5
	Total	18 / 20 Passed

Validation

68%

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 — 11 / 16 Passed

Validation for skill structure

Criteria	Description	Result
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
metadata_version	'metadata.version' is missing	Warning
metadata_field	'metadata' should map string keys to string values	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning
referenced_paths_exist	Referenced path issues: 1 missing	Warning

	Total	11 / 16 Passed

Repository: JaviMontano/claude-plugins
Path: plugins/claude-native-toolkit/skills/agent-harness-diagnose/SKILL.md
Commit: e8f986b

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.