he-reinforce
Create or refresh evidence-bound Harness Engineering learning artifacts from verified solved problems. Use when a fix worked, a repeated failure should become durable knowledge, or .harness/solutions and Project Brain need maintenance.
46
47%
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/skills/he-reinforce/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 does well at establishing a clear niche and answering both 'what' and 'when' with explicit trigger conditions. However, the specific capabilities could be more concretely enumerated (e.g., what exactly does 'refresh' entail?), and the trigger terms lean heavily on internal jargon rather than natural user language. The domain-specific terminology helps distinctiveness but may hurt discoverability if users phrase requests differently.
Suggestions
List more concrete actions beyond 'create or refresh' — e.g., 'document root causes, update solution files, link evidence to failure patterns'
Add natural-language trigger variations users might say, such as 'save this solution', 'record what fixed this', 'update knowledge base'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names a domain ('Harness Engineering learning artifacts') and some actions ('create or refresh'), but the concrete actions are not comprehensively listed. Terms like 'evidence-bound' and 'learning artifacts' are somewhat abstract rather than describing specific operations. | 2 / 3 |
Completeness | The description clearly answers both 'what' (create or refresh evidence-bound learning artifacts from verified solved problems) and 'when' (when a fix worked, when a repeated failure should become durable knowledge, or when .harness/solutions and Project Brain need maintenance) with explicit trigger conditions. | 3 / 3 |
Trigger Term Quality | Includes some relevant keywords like 'fix worked', 'repeated failure', '.harness/solutions', 'Project Brain', and 'durable knowledge', but these are fairly niche/internal terms. A user might say 'save this fix' or 'update solutions' but the description doesn't cover those natural variations well. | 2 / 3 |
Distinctiveness Conflict Risk | The description is highly specific to a particular system (.harness/solutions, Project Brain, Harness Engineering) with distinct triggers that are unlikely to conflict with other skills. The niche is clearly defined. | 3 / 3 |
Total | 10 / 12 Passed |
Implementation
20%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is heavily abstract and framework-internal, reading more like an architecture specification than actionable guidance for Claude. It lacks concrete examples, executable code, and clear step-by-step instructions, instead relying on jargon-heavy descriptions and repeated references to folded context files. The workflow has some structure but is incomplete and would benefit significantly from concrete examples showing actual inputs and outputs for each mode.
Suggestions
Add concrete, executable examples for at least the primary mode (capture_solved_problem) showing actual input evidence, the commands/steps to run, and the expected output artifact with real frontmatter and content.
Complete the truncated procedure step 4 and convert the abstract procedure into a concrete numbered workflow with specific commands, file paths, and validation checkpoints.
Remove or drastically reduce the abstract descriptions of modes, statuses, and output fields — instead show one complete worked example that demonstrates the full flow from evidence to artifact.
Consolidate the repeated 'See references/hot-path-folded-context.md for folded X detail' pattern into a single reference section entry rather than scattering it across every section.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose with heavy jargon, abstract descriptions, and repeated references to folded context files. It explains internal framework concepts at length without providing any concrete, executable guidance. Many sections are padded with lists of statuses and modes that don't add actionable value. | 1 / 3 |
Actionability | There is no concrete, executable code, no specific commands (except one script path buried in Validation), no copy-paste-ready examples, and no clear step-by-step instructions. The 'Examples' section describes scenarios abstractly rather than showing actual inputs and outputs. The procedure is vague and incomplete (step 4 is cut off). | 1 / 3 |
Workflow Clarity | The procedure section attempts to outline a multi-step process with mode selection, eligibility proving, scope limiting, and validation, but steps are abstract and incomplete (step 4 is truncated). Validation mentions a specific script and pass/fail/blocked gates, which is good, but the feedback loop for error recovery is only loosely described in Failure Handling rather than integrated into the workflow. | 2 / 3 |
Progressive Disclosure | The skill references multiple external files (contract.yaml, evals.yaml, hot-path-folded-context.md, etc.) which shows an attempt at progressive disclosure. However, no bundle files were provided to verify these exist, multiple sections end with 'See references/hot-path-folded-context.md for folded X detail' which feels like a crutch rather than clear navigation, and the main content itself is still quite long despite offloading detail. | 2 / 3 |
Total | 6 / 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
d00c351
- 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.