he-router
Route ambiguous Harness Engineering requests to one lifecycle stage when users ask where to start, resume, plan, implement, review, debug, schedule a heartbeat, or resolve domain terminology.
62
73%
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 ./Infrastructure/references/deferred-skill-context/harness-engineering-he-router/SKILL.mdQuality
Discovery
89%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 is a solid description that clearly communicates its routing purpose within the Harness Engineering domain. It provides explicit trigger conditions and domain-specific terminology that make it distinctive. The main weakness is that the core capability ('route to one lifecycle stage') is somewhat abstract—it could benefit from slightly more specificity about what the routing output looks like.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names a domain ('Harness Engineering requests') and lists several actions (start, resume, plan, implement, review, debug, schedule a heartbeat, resolve domain terminology), but the core action is 'route to one lifecycle stage' which is somewhat abstract. The listed actions are more like trigger scenarios than concrete capabilities the skill performs. | 2 / 3 |
Completeness | The description clearly answers both 'what' (route ambiguous Harness Engineering requests to one lifecycle stage) and 'when' (when users ask where to start, resume, plan, implement, review, debug, schedule a heartbeat, or resolve domain terminology). The trigger conditions are explicitly stated. | 3 / 3 |
Trigger Term Quality | Good coverage of natural terms users would say: 'where to start', 'resume', 'plan', 'implement', 'review', 'debug', 'schedule a heartbeat', and 'resolve domain terminology' are all plausible user phrases. The inclusion of 'ambiguous' also helps clarify the routing context. | 3 / 3 |
Distinctiveness Conflict Risk | The description is highly specific to 'Harness Engineering' and 'lifecycle stage' routing, with domain-specific terms like 'heartbeat' that create a clear niche. It is unlikely to conflict with other skills due to its specialized routing function. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
57%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-organized routing skill with strong progressive disclosure and clear reference structure. Its main weaknesses are the lack of concrete output examples (no sample JSON showing the expected structured response) and some verbosity in procedure steps that could be tightened. The workflow would benefit from inline validation checkpoints and at least one complete input→output example.
Suggestions
Add at least one complete input→output example showing the full structured response (selected_stage, matched_rule, confidence, rationale, next_invocation, subagent policy) so Claude knows the exact output format.
Remove the 'because' explanatory clauses from procedure steps (e.g., step 1 and 2) since Claude can infer the reasoning from the rule structure itself.
Integrate validation gates inline within the procedure (e.g., after step 3: 'If no rule matched, go to step 10') rather than having a separate Validation section at the end.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is moderately efficient but includes some redundancy—several routing rules in the Procedure section overlap with Anti Patterns and Constraints, and the Examples section lists natural-language prompts without showing expected outputs, adding bulk without proportional value. Some explanatory 'because' clauses in procedure steps are unnecessary for Claude. | 2 / 3 |
Actionability | The procedure provides a clear deterministic decision sequence and the outputs section specifies structured fields, but there are no concrete output examples (e.g., a sample JSON response with selected_stage, rationale, next_invocation). The examples section lists input prompts but never shows the expected routing output, leaving Claude to infer the exact response format. | 2 / 3 |
Workflow Clarity | The 10-step procedure is sequenced and includes a 'fail fast at first failed gate' validation note, but there are no explicit checkpoints between steps (e.g., 'confirm artifact state before proceeding to step 4') and the validation section is separated from the workflow rather than integrated as inline gates. The 'if still ambiguous' fallback in step 10 is good but the intermediate steps lack similar feedback loops. | 2 / 3 |
Progressive Disclosure | The skill is well-structured as a routing overview with clearly signaled one-level-deep references to external files (routing-map.json, deterministic-stage-routing.md, session-evidence-contract.md, etc.), each annotated with 'Read when' conditions. Content is appropriately split between the overview and referenced materials, and navigation is easy. | 3 / 3 |
Total | 9 / 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
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.