he-compound
Run a bounded Harness Engineering lifecycle across multiple stages. Use when the user wants coordinated brainstorm, spec, plan, work, review, and fix flow rather than one isolated stage.
48
51%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./Plugins/harness-engineering/fixtures/budget-archive/2026-04-21/deferred-store/skills/team_automation/he-compound/SKILL.mdQuality
Discovery
75%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 is well-structured with a clear 'Use when' clause that distinguishes this multi-stage coordination skill from single-stage skills. Its main weaknesses are the use of domain-specific jargon ('Harness Engineering') that users may not naturally use, and the lack of concrete action descriptions beyond listing stage names. Overall it's functional but could benefit from more natural trigger terms and slightly more specificity about what each stage entails.
Suggestions
Add natural trigger terms users might say, such as 'workflow', 'end-to-end', 'full process', 'multi-step development', or 'project lifecycle'.
Briefly describe concrete actions for at least the key stages (e.g., 'generates specifications, creates implementation plans, executes code changes, runs reviews') rather than just listing stage names.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (Harness Engineering lifecycle) and lists the stages (brainstorm, spec, plan, work, review, fix), but doesn't describe concrete actions within each stage—just names them as a flow. | 2 / 3 |
Completeness | Clearly answers both 'what' (run a bounded Harness Engineering lifecycle across multiple stages) and 'when' (when the user wants coordinated brainstorm, spec, plan, work, review, and fix flow rather than one isolated stage), with an explicit 'Use when' clause and a distinguishing contrast ('rather than one isolated stage'). | 3 / 3 |
Trigger Term Quality | Includes some relevant terms like 'brainstorm', 'spec', 'plan', 'review', 'fix', and 'lifecycle', but 'Harness Engineering' is jargon that users may not naturally say. Missing common variations like 'workflow', 'pipeline', 'end-to-end', 'multi-step', or 'full process'. | 2 / 3 |
Distinctiveness Conflict Risk | The description carves out a clear niche—coordinated multi-stage lifecycle flow—and explicitly distinguishes itself from skills that handle a single isolated stage, making it unlikely to conflict with individual-stage skills. | 3 / 3 |
Total | 10 / 12 Passed |
Implementation
27%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill reads as a high-level orchestration policy document rather than an actionable skill. It describes what should happen at an abstract level but never shows Claude how to do it — no concrete commands, no file paths to inspect, no example outputs, and no linked reference documents for the stages and tools it names. The workflow has reasonable structure but lacks the specificity needed for Claude to execute reliably.
Suggestions
Add concrete, executable examples: show exactly what Claude should do at each stage (e.g., specific file paths to check for artifacts, exact commands to run, example output format with schema_version: 1).
Link to or inline documentation for referenced concepts like 'skill-refactor', 'skillify', 'he-compound-refresh', and the individual Harness Engineering stages — currently these are opaque references with no navigation.
Replace abstract procedure steps with specific decision trees or checklists (e.g., 'Check for docs/solutions/*.md matching the topic → if found and overlap > 80%, refresh; otherwise create new').
Add at least one complete input→output example showing the full flow from request to structured output with schema_version: 1, result, validation, blockers, and next action fields.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is moderately efficient but includes philosophical framing and abstract descriptions that don't add actionable value. Phrases like 'Progressive-disclosure entrypoint for stage orchestration and durable learning capture' are jargon-heavy without being instructive. Some sections (e.g., 'When to use') partially repeat the description. | 2 / 3 |
Actionability | The skill is almost entirely abstract guidance with no concrete commands, code, executable examples, or specific tool invocations. Instructions like 'Select lifecycle mode using artifact-first evidence' and 'continue from the earliest incomplete or untrusted stage' describe intent without telling Claude exactly what to do, what files to check, or what commands to run. | 1 / 3 |
Workflow Clarity | The Procedure section provides a numbered sequence and the Validation section includes a 'fail fast' gate, which is good. However, the steps are abstract rather than concrete (e.g., 'Select lifecycle mode using artifact-first evidence' — how?), validation checkpoints lack specifics on what constitutes passing/failing, and there are no feedback loops for error recovery beyond 'stop at first failed gate.' | 2 / 3 |
Progressive Disclosure | The skill references several external concepts (stages like brainstorm/spec/plan/work/review, 'skill-refactor', 'skillify', 'he-compound-refresh', 'docs/solutions/') but provides no links to documentation for any of them. No bundle files are provided. The 'Full Context' section only links to icon images, which are irrelevant to the skill's purpose. Navigation to deeper content is effectively absent. | 1 / 3 |
Total | 6 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
metadata_version | 'metadata.version' is missing | Warning |
Total | 10 / 11 Passed | |
- 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.