he-fix-bugs

Debug and repair validated Harness Engineering defects with bounded scope, reproduction evidence, root-cause notes, regression protection, and validation proof. Use when a bug is already evidenced and the fix should not expand into broad improvement work.

Quality

80%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Failed to scan

The risk profile of this skill

Fix and improve this skill with Tessl

tessl review fix ./Plugins/harness-engineering/skills/he-fix-bugs/SKILL.md

Quality

Content

75%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 well-structured, competent bug-fix procedure with explicit validation feedback loops and a concrete output schema. It is held back from higher scores by mild redundancy, abstract (non-executable) validation commands, a duplicated step number, and references to unverifiable paths outside the package.

Suggestions

Fix the duplicated step '6' in the Procedure and clarify the ordering of validate → stage → store-media.

Give the validation gates executable forms (e.g., actual command invocations) rather than only naming them.

Either bring the externally-referenced contracts (stage-arc-boundary-contract.md, subagent-call-contract.md, visual-reference-contract.md) into references/ or confirm they resolve, so navigation is reliable.

De-duplicate the approval/side-effect guidance that recurs across Preconditions, Safety Boundaries, and Failure Handling.

Dimension	Reasoning	Score
Conciseness	The body is lean and procedural and assumes Claude's competence without explaining basics, but approval language is repeated across Preconditions, Safety Boundaries, and Failure Handling, and the Procedure has a duplicated step '6' — minor instances that could be trimmed.	4 / 5
Actionability	Concrete procedural steps and a specific output schema (schema_version: 1, side-effect class, pass\|fail\|blocked, named stage_arc_boundary lenses) are given, but validation gates are named abstractly ('strict audit, OpenClaw, OpenAI format lint') rather than as executable commands, leaving a minor gap.	4 / 5
Workflow Clarity	An 8-step sequence with an explicit 'Fail fast' validation feedback loop (stop → fix/block → rerun) and checklists is present, satisfying the destructive-op validation requirement, but the duplicated step '6' numbering and muddled step-5/6 ordering are minor structural gaps.	4 / 5
Progressive Disclosure	Well-organized sections with one-level-deep local references that exist (references/contract.yaml, references/evals.yaml), but several references point outside the package (../../references/stage-arc-boundary-contract.md, ../../references/subagent-call-contract.md, etc.) and cannot be verified, creating a navigation risk.	4 / 5
	Total	16 / 20 Passed

Description

85%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 with clear what-and-when structure and a distinct niche. Its main weakness is trigger-term quality: the natural keywords sit inside heavy domain jargon and several common synonyms are absent.

Suggestions

Add common user-facing synonyms in the 'Use when' clause (e.g., 'a failing test, broken behavior, error, or already-evidenced bug').

Soften or briefly gloss the in-house jargon ('Harness Engineering defects', 'regression protection', 'validation proof') so the trigger reads more like natural user speech.

Consider naming the artifact types users reference (test failures, CI failures, stack traces) as trigger terms.

Dimension	Reasoning	Score
Specificity	Names the domain and lists multiple concrete actions and deliverables — 'Debug and repair', 'reproduction evidence', 'root-cause notes', 'regression protection', 'validation proof' — giving comprehensive, non-vague coverage.	5 / 5
Completeness	Explicitly states what it does ('Debug and repair validated Harness Engineering defects...') and when to use it ('Use when a bug is already evidenced and the fix should not expand into broad improvement work'), with concrete trigger phrasing.	5 / 5
Trigger Term Quality	Natural terms 'bug', 'fix', 'debug', and 'repair' appear, but they are embedded in dense domain jargon ('Harness Engineering defects', 'regression protection', 'validation proof') and common synonyms a user would say (error, broken, failing, crash) are missing.	3 / 5
Distinctiveness Conflict Risk	The 'validated Harness Engineering defects' + 'bounded scope' + 'should not expand into broad improvement work' framing carves a clear niche with distinct triggers and minimal overlap risk.	5 / 5
	Total	18 / 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: jscraik/Agent-Skills
Path: Plugins/harness-engineering/skills/he-fix-bugs/SKILL.md
Commit: 2f91362

Reviewed: 1 day 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.