agent-reproduce-feature

Use when reproducing an existing Codex or Claude Code feature in Qwen Code or another agent CLI by choosing a reference agent, capturing HTTP request bodies, prompts, tool/function schemas, terminal output, and then implementing the matching behavior in the target repo.

Quality

86%

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

87%

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

A well-structured, action-dense skill body with executable commands and clean one-level-deep references. The main gap is workflow gating: validation checkpoints exist (checklist, TLS recovery) but are not embedded as explicit pass/fail gates within the capture-to-implementation sequence.

Suggestions

Insert an explicit in-flow validation gate after capture, e.g. step 5a: 'Confirm the trace contains the feature's request body and tool schemas before implementing; if TLS failed or traffic is missing, re-run capture per references/capture-workflow.md before proceeding.'

Mark Done Criteria as a gated checklist referenced from the workflow (e.g. 'Do not hand off until every Done Criteria item is verified') so the verification loop is sequenced rather than only appended.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes competence — it never explains what Codex/Claude Code/Qwen Code are, and every section is action-oriented commands or directives with no concept padding, matching the 'every token earns its place' anchor rather than the score-2 case that retains filler justification.	3 / 3
Actionability	It provides fully executable, copy-paste-ready commands — command -v discovery, mkdir, run_with_mitm.sh / run_tmux_capture.sh / capture_state.py invocations with real flags and args — matching the fully-executable anchor rather than the pseudocode score-2 anchor.	3 / 3
Workflow Clarity	There is a clear 8-step sequence, a Done Criteria checklist, and a TLS error-recovery note, but the in-flow validation checkpoints for the fragile capture step (e.g. 'confirm the trace contains the feature request body before implementing') are implicit rather than gated within the sequence, fitting 'sequence present but checkpoints implicit' over the inline-gated score-3 anchor.	2 / 3
Progressive Disclosure	The body is an overview pointing one level deep to a real references/capture-workflow.md (explicitly signaled: 'Read ... before running capture for the first time') and real executable scripts, with content appropriately split, matching the well-signaled one-level-deep anchor; all referenced paths resolve to existing bundle files.	3 / 3
	Total	11 / 12 Passed

Description

85%

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 strong, specific description with an explicit 'Use when' trigger and a clear, distinctive niche. Its main weakness is trigger-term breadth: it captures the canonical phrasing but omits common synonyms a user might naturally say.

Suggestions

Add common phrasing variations to the trigger, e.g. 'Use when porting, mirroring, or reproducing a Codex or Claude Code feature in Qwen Code or another agent CLI', so the skill matches users who say 'port' or 'reimplement' rather than 'reproduce'.

Dimension	Reasoning	Score
Specificity	It lists multiple concrete actions — 'choosing a reference agent, capturing HTTP request bodies, prompts, tool/function schemas, terminal output, and then implementing the matching behavior' — matching the multi-action anchor rather than the single-domain score-2 anchor.	3 / 3
Completeness	It explicitly answers both what (choose reference agent, capture bodies/prompts/schemas/output, implement matching behavior) and when via the explicit 'Use when reproducing...' clause, matching the score-3 what-and-when anchor; it is not the score-2 case because the when trigger is explicit, not implied.	3 / 3
Trigger Term Quality	The core trigger 'reproducing an existing Codex or Claude Code feature in Qwen Code or another agent CLI' is natural, but it misses common variations a user might equally say (port, mirror, reimplement, copy behavior), so it sits at 'some relevant keywords but missing common variations' rather than full coverage.	2 / 3
Distinctiveness Conflict Risk	The niche — reproducing a Codex/Claude Code feature in Qwen Code via HTTP/state capture — is highly specific with distinct triggers unlikely to fire for unrelated skills, matching the clear-niche anchor rather than the overlapping score-2 anchor.	3 / 3
	Total	11 / 12 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: QwenLM/qwen-code
Path: .qwen/skills/agent-reproduce-feature/SKILL.md
Commit: 62e009a

Reviewed: 5 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.