CtrlK
BlogDocsLog inGet started
Tessl Logo

harness-loop

Long-running task loop using /loop (Claude Code dynamic mode) and ScheduleWakeup to re-enter with fresh context on each wake-up. Internally invokes harness-work through Agent. Trigger: long-running, loop, wake-up, autonomous. Do NOT load for: one-shot task execution, review, release, planning.

65

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./opencode/skills/harness-loop/SKILL.md

The canonical home for this skill is harness-loop in Chachamaru127/claude-code-harness

SKILL.md
Quality
Evals
Security

Quality

Content

71%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A detailed, well-structured operational spec with concrete commands, explicit validation checkpoints, and a clean one-level reference to flow.md. Its main weakness is mild verbosity and some duplication between the inline flow summary and the referenced detailed flow.

Suggestions

Trim redundant prose: state the [60,3600] clamp constraint once (in the pacing note) rather than re-explaining each value's validity in prose.

Reduce overlap between the inline wake-up flow and references/flow.md — keep SKILL.md to the high-level sequence with step names and defer all command details to flow.md.

Cut the analogy asides ('たとえると、作業机…工具箱…') that explain concepts Claude already understands; state the rule directly.

DimensionReasoningScore

Conciseness

Mostly efficient with tables and a compact ASCII flow diagram, but includes some unnecessary repetition (the [60,3600] clamp and pacing rationale restated in prose, the wake-up flow summarized inline yet also deferred to flow.md) and explanatory analogies that could be trimmed.

3 / 5

Actionability

Provides concrete executable commands throughout (generate-sprint-contract.js, ensure-sprint-contract-ready.sh, git cherry-pick, ScheduleWakeup with full params, Agent() call), with only minor gaps where steps delegate to flow.md.

4 / 5

Workflow Clarity

A clearly sequenced 10-step wake-up flow with explicit validation checkpoints (contract readiness, APPROVE/REQUEST_CHANGES review loop with max_iterations, plateau exit codes, cycle cap) and tabulated stop conditions — appropriate feedback loops for this batch/destructive operation.

5 / 5

Progressive Disclosure

SKILL.md is a well-sectioned overview with a single one-level reference to references/flow.md (verified to exist) signaled via the full CLAUDE_SKILL_DIR path; minor organization gap from the wake-up flow being inlined in SKILL.md and expanded again in flow.md.

4 / 5

Total

16

/

20

Passed

Description

86%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A specific, well-scoped description that states both what the skill does and when to load it, reinforced by an explicit exclusion list. Trigger terms are natural and the niche is clearly distinguishable from related harness skills.

DimensionReasoningScore

Specificity

Lists several concrete actions ('Long-running task loop using /loop', 'ScheduleWakeup to re-enter with fresh context', 'Internally invokes harness-work through Agent') naming both mechanism and sub-actions; falls just short of comprehensive feature coverage.

4 / 5

Completeness

Explicitly answers both what (loop mechanism, fresh-context re-entry, harness-work via Agent) and when (concrete trigger phrases plus an explicit 'Do NOT load for' exclusion list).

5 / 5

Trigger Term Quality

Provides natural user-facing terms ('Trigger: long-running, loop, wake-up, autonomous') that a user would plausibly say, though synonym/extension coverage is a bit thin.

4 / 5

Distinctiveness Conflict Risk

Highly specific to the /loop + ScheduleWakeup mechanism, and the explicit 'Do NOT load for: one-shot task execution, review, release, planning' actively disambiguates from sibling skills like harness-review and harness-plan.

5 / 5

Total

18

/

20

Passed

Validation

93%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 1 missing, 1 deeper-than-1-level

Warning

Total

15

/

16

Passed

Repository
Chachamaru127/claude-code-harness
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.