he-deepen-plan
Deepen an existing implementation plan so sequencing, verification, and risk treatment are strong enough for execution. Use when the user wants Harness Engineering plan hardening before he-work.
43
43%
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-deepen-plan/SKILL.mdQuality
Discovery
52%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 attempts to cover both 'what' and 'when' but is undermined by poor trigger term quality, apparent typos ('he-work'), and domain-specific jargon ('Harness Engineering plan hardening') that users are unlikely to naturally use. The specificity of capabilities is moderate but would benefit from listing concrete actions rather than a single abstract verb ('deepen').
Suggestions
Replace jargon and apparent typos ('Harness Engineering plan hardening', 'he-work') with natural language users would actually say, e.g., 'Use when the user wants to strengthen, refine, or harden an implementation plan before starting development work.'
List specific concrete actions the skill performs, e.g., 'Reviews task sequencing and dependencies, adds verification checkpoints, identifies risks and mitigation strategies, and ensures the plan is execution-ready.'
Add natural trigger terms users might say, such as 'refine plan', 'plan review', 'improve implementation plan', 'execution readiness', 'risk assessment'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names a domain ('implementation plan') and a general action ('deepen...so sequencing, verification, and risk treatment are strong enough for execution'), but it doesn't list multiple concrete actions—it's more of a high-level summary of what 'deepening' entails rather than specific discrete operations. | 2 / 3 |
Completeness | The description does explicitly answer both 'what does this do' (deepen an existing implementation plan focusing on sequencing, verification, and risk treatment) and 'when should Claude use it' ('Use when the user wants Harness Engineering plan hardening before he-work'). However, the 'when' clause is poorly worded and contains apparent errors. | 3 / 3 |
Trigger Term Quality | The trigger terms are weak and unnatural. 'Harness Engineering plan hardening' and 'he-work' are jargon or possibly typos that users would not naturally say. Common terms like 'refine plan', 'improve plan', 'plan review', 'execution readiness' are absent. | 1 / 3 |
Distinctiveness Conflict Risk | The description is somewhat specific to plan hardening/deepening, which narrows the niche, but the vague jargon ('Harness Engineering plan hardening', 'he-work') makes it unclear what exactly distinguishes this from other planning or project management skills. | 2 / 3 |
Total | 8 / 12 Passed |
Implementation
35%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 process description rather than actionable operational guidance. While it has reasonable structure and references external files for progressive disclosure, the core procedure is too abstract to be executable—it tells Claude to 'deepen' and 'inspect' without specifying how. For a skill focused on plan hardening, it ironically lacks the concrete checkpoints and validation criteria it advocates.
Suggestions
Add a concrete template or structured format for the deepened plan output (e.g., a markdown table with columns for Task, Dependencies, Gate Criteria, Rollback Steps, Owner) so Claude knows exactly what to produce.
Replace the abstract procedure steps with specific, actionable instructions—e.g., 'For each task in the plan, verify it has: (1) explicit inputs and outputs, (2) a named validation method, (3) a rollback procedure. If any are missing, add them.'
Include at least one worked example showing a before/after of a thin plan being deepened, demonstrating the expected transformation concretely.
Link or inline the 'archived full guide' mentioned in the Progressive Disclosure Entry section—currently it is referenced but never provided or linked.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is moderately efficient but includes some unnecessary sections (Philosophy, Anti-patterns) and redundant phrasing. The Subagent Routing section is verbose with details about manifest resolution and agent creation that could be more tightly expressed. However, it avoids explaining basic concepts Claude already knows. | 2 / 3 |
Actionability | The skill provides only abstract, high-level guidance with no concrete code, commands, or executable examples. Instructions like 'Inspect current plan for ambiguity, hidden coupling, and missing checkpoints' and 'Deepen task sequencing' are vague directives without specific techniques, templates, or structured output formats to follow. | 1 / 3 |
Workflow Clarity | The procedure lists three steps in sequence and the validation section mentions a fail-fast gate, but the steps themselves are abstract ('inspect', 'deepen', 'return') without concrete checkpoints or feedback loops. For a skill that explicitly aims to add verification and risk treatment to plans, the workflow itself lacks the rigor it prescribes. | 2 / 3 |
Progressive Disclosure | The skill references multiple external files (contract.yaml, evals.yaml, task-profile.json, subagent-routing.md) with clear links, and mentions an 'archived full guide' for detailed policy. However, no bundle files were provided to verify these references exist, the 'archived full guide' is never linked, and the references section mixes canonical and non-canonical sources in a way that could confuse navigation. | 2 / 3 |
Total | 7 / 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.