terminal-bench-loop

Run one Terminal-Bench task through a bounded Paperclip smoke/diagnosis/fix loop. Use when asked to drive Terminal-Bench until it passes, rerun a Terminal-Bench loop, or iterate with board-gated fixes and diagnosis.

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

77%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.

The content is a well-structured operational contract with excellent workflow sequencing, validation checkpoints, and a verification checklist, and it respects Claude's intelligence without padding. Its main weakness is progressive disclosure: the single referenced file is absent from the bundle and the document is monolithic.

Suggestions

Add the referenced `doc/execution-semantics.md` to the bundle (or change the reference to a path that exists) so the one-level-deep pointer resolves instead of dangling.

Split the long inlined sections (issue topology, liveness rule, pitfalls) into clearly-signaled reference files under `references/` with concise pointers from SKILL.md to reduce the monolithic wall.

Tighten the repeated issue-state transition prose in Steps 6 and 8, which restate the same `in_review`-vs-`blocked` distinction several times.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes Claude's competence — it never explains what Terminal-Bench, Paperclip, or Harbor are — but it is a long (~230 lines) single document with some elaboration around issue-state transitions that could be tightened, keeping it just below the lean anchor.	4 / 5
Actionability	Provides concrete, executable guidance — specific commands (`pnpm smoke:terminal-bench-loop-skill`), env vars (`PAPERCLIPAI_CMD`, `PAPERCLIP_HARBOR_RUNNER_CONFIG`), idempotency keys, and exact status transitions — with only minor gaps since the bulk is instruction/control-plane rather than copy-paste code.	4 / 5
Workflow Clarity	A 9-step procedure with explicit validation checkpoints, a dedicated 'Verification checklist', and clear feedback loops (validate -> fix -> retry, board rejection -> stop), matching the anchor for clear sequencing with explicit validation and error-recovery loops.	5 / 5
Progressive Disclosure	The body references `doc/execution-semantics.md` as a one-level-deep source of truth, but no `doc/`, `references/`, `scripts/`, or `assets/` bundle directories exist, so the reference is dangling; the ~230-line monolith also inlines topology, liveness, and pitfall content that could live in separate files.	3 / 5
	Total	16 / 20 Passed

Description

87%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.

The description cleanly states a specific operational domain and pairs a concrete 'what' with an explicit 'Use when' trigger clause, satisfying both completeness and distinctiveness well. Trigger-term and specificity are strong with only minor coverage gaps.

Dimension	Reasoning	Score
Specificity	Names the domain ('Terminal-Bench task', 'Paperclip') and lists several concrete actions — running a task through a 'smoke/diagnosis/fix loop' — but does not enumerate the full set of sub-actions (board-gated fixes, worktree continuity) beyond naming them, leaving minor coverage gaps.	4 / 5
Completeness	Explicitly answers both 'what' (run one Terminal-Bench task through a bounded smoke/diagnosis/fix loop) and 'when' (a 'Use when' clause listing concrete trigger situations), matching the anchor for clearly answering both with concrete triggers.	5 / 5
Trigger Term Quality	Includes natural trigger phrases users would say ('drive Terminal-Bench until it passes', 'rerun a Terminal-Bench loop', 'iterate with board-gated fixes'), with good keyword coverage, but the terms are proper-noun heavy and miss some plain synonyms a user might casually utter.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear niche (Terminal-Bench loop driving) with distinct triggers and explicit out-of-scope boundaries implied by the operational/diagnostic framing, giving minimal conflict risk with other skills.	5 / 5
	Total	18 / 20 Passed

Validation

100%

Warnings & errors only

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: paperclipai/paperclip
Path: .agents/skills/terminal-bench-loop/SKILL.md
Commit: af3023f

Reviewed: 4 days ago

Table of Contents

Discovery Implementation Validation

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.