he-phase-work
Runs approved Harness Engineering plans in recurring phases: check live state, wake with he-heartbeat, execute only the active he-work slice, verify gates, update Linear only with approval, and stop before unsafe staging or closure. Use when a plan, issue, or PR needs continued phase-by-phase execution.
68
81%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
85%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This description is strong in specificity and completeness, clearly listing concrete actions and providing an explicit 'Use when' clause. The domain-specific jargon makes it highly distinctive but may reduce trigger term quality since users might not use terms like 'he-heartbeat' or 'he-work slice' naturally. Overall it serves its purpose well for a specialized internal workflow skill.
Suggestions
Add more natural user-facing trigger terms alongside the jargon, e.g., 'Use when a plan, issue, PR, or workflow needs continued phase-by-phase execution, or when resuming Harness Engineering work.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: check live state, wake with he-heartbeat, execute active he-work slice, verify gates, update Linear with approval, and stop before unsafe staging or closure. These are detailed, actionable steps. | 3 / 3 |
Completeness | Clearly answers both what (runs approved Harness Engineering plans through recurring phases with specific steps) and when ('Use when a plan, issue, or PR needs continued phase-by-phase execution'), with an explicit trigger clause. | 3 / 3 |
Trigger Term Quality | Includes some relevant terms like 'plan', 'issue', 'PR', 'phase-by-phase execution', 'Linear', and 'Harness Engineering', but uses domain-specific jargon ('he-heartbeat', 'he-work slice', 'gates') that users may not naturally say. Missing common variations of how users might request this. | 2 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with domain-specific terminology like 'Harness Engineering plans', 'he-heartbeat', 'he-work slice', and 'Linear'. Very unlikely to conflict with other skills due to its narrow, specialized niche. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured orchestration skill with strong workflow clarity, explicit validation gates, and concrete delegation examples. Its main weaknesses are moderate verbosity in listing inputs/constraints and the inability to verify referenced files exist in the bundle. The actionability is excellent with executable commands, specific examples, and a structured output template.
Suggestions
Tighten the 'Inputs' section into a compact table or inline list rather than a long comma-separated enumeration to improve conciseness.
Verify that referenced files (references/phase-gate-contract.md, references/contract.yaml, etc.) exist in the bundle, or remove references that cannot be resolved.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably efficient but includes some sections that could be tightened—e.g., the 'When Not to Use' and 'Inputs' sections list many items that could be more compact. The philosophy section and some constraint restatements add mild verbosity, though overall it respects Claude's intelligence. | 2 / 3 |
Actionability | The skill provides concrete executable validation commands (git status, rg, test -f, audit script), specific delegation examples with exact commands and file paths, and a structured YAML output template. The procedure steps are specific and copy-paste ready. | 3 / 3 |
Workflow Clarity | The 7-step procedure is clearly sequenced with explicit validation checkpoints (step 5), fail-fast policy, feedback loops (route failures to he-work or he-code-review), and explicit gates requiring approval before destructive actions (step 7). The failure mode section reinforces stopping at the first failed gate. | 3 / 3 |
Progressive Disclosure | References to external files (phase-gate-contract.md, contract.yaml, subagent-call-contract.md) are listed but no bundle files were provided to verify they exist. The skill itself is somewhat long and could benefit from splitting delegation patterns or the output template into separate reference files. The references section is present but not deeply signaled throughout the document. | 2 / 3 |
Total | 10 / 12 Passed |
Validation
100%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 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- jscraik/Agent-Skills
- Commit
4c78f98
- Reviewed
Table of Contents
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.