ugc-fixloop

the UGC fix-loop toolkit — surgically re-render a bad window/beat of a single-take UGC master (stitch_replacement.py, pure FFmpeg) and GPT cross-model review a Seedance prompt before render (vet_seedance_prompt.py, routed through the openai-proxy). Fetch it into a one-shot UGC recipe so both scripts resolve on any machine and the vet call bills the Ads agent.

Quality

84%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

93%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 SKILL.md body is an exemplary lean operational doc: complete executable CLI references with defaults, explicit validation checkpoints (dry-run, duration-drift warning, exit-3 fallback), correct dependency notes, and a bundle structure that matches what the body claims. The only notable gap is that the overall fix-loop sequence across the two tools is implied rather than presented as an explicit ordered workflow.

Dimension	Reasoning	Score
Conciseness	The body is dense and lean with zero padding — no concept explanations, just operational facts ('no API key, no network', 'the master's audio (VO + ambience) plays straight through, so lip-sync on talking beats is never touched'). Every token carries information Claude doesn't already know, matching anchor 5 rather than anchor 4, where something could be trimmed.	5 / 5
Actionability	Fully executable guidance: complete command lines for both scripts, every flag documented with defaults ('--scene-threshold 0.3', '--fit {stretch,trim,freeze}'), the exact proxy endpoint path, the credentials file location, and exit-code semantics ('Exits 3 if the proxy/creds are unavailable'). Copy-paste ready and covers both common window-selection cases, matching anchor 5.	5 / 5
Workflow Clarity	Validation checkpoints are explicitly present — '--dry-run — print the ffmpeg command without running', 'Warns if output duration drifts >0.15s from the master (audio-sync check)', and the exit-3 advisory-not-gate fallback — so the destructive-operation cap at 3 does not apply (and the master is never modified; output is a separate file). It sits at anchor 4 rather than 5 because the end-to-end fix-loop sequence (vet prompt → render replacement via create-video-fal → stitch → check drift warning) is implied by the sections rather than laid out as an explicit ordered flow.	4 / 5
Progressive Disclosure	The body is under 50 lines with well-organized sections, and the bundle structure matches the body: both scripts it names ('stitch_replacement.py', 'vet_seedance_prompt.py') exist in scripts/ with no nested references and no content that belongs in a separate file. The simple-skill exception applies, so well-organized sections with the actual bundle backing the named scripts merit 5.	5 / 5
	Total	19 / 20 Passed

Description

70%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 is highly specific, concrete, and unambiguous about what the toolkit does and how it integrates (proxy routing, billing, fetch location), with essentially no conflict risk against other skills. Its main weakness is the missing explicit 'when to use' trigger clause, and secondarily a reliance on internal jargon that limits natural trigger phrases a user might say.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when a one-shot UGC video recipe produces a take with a bad beat/window that needs a surgical re-render, or before rendering a Seedance prompt you want cross-model reviewed.'

Include natural trigger phrases/synonyms users would actually say, such as 'fix a bad clip', 'replace a segment/scene of a UGC video', or 're-render one beat'.

Trim internal operational jargon ('bills the Ads agent', 'one-shot UGC recipe') or rephrase it in user-facing terms so the trigger terms read naturally.

Dimension	Reasoning	Score
Specificity	Names multiple concrete actions ('surgically re-render a bad window/beat of a single-take UGC master', 'GPT cross-model review a Seedance prompt before render', 'Fetch it into a one-shot UGC recipe') with the implementing scripts and transport details. It falls just short of anchor 5's comprehensive coverage because the actions are wrapped in heavy internal jargon, but is clearly above anchor 3's 1-2 concrete actions.	4 / 5
Completeness	The 'what' is clear (two scripts, what each does, how they're deployed and billed), but there is no 'Use when...' clause or equivalent explicit trigger guidance — 'before render' only weakly implies when. Per the judging guidelines, a missing explicit trigger clause caps completeness at 3.	3 / 5
Trigger Term Quality	Includes relevant keywords ('UGC', 're-render', 'Seedance prompt', 'FFmpeg', 'fix-loop') that a user of this domain would say. Natural variations like 'fix a bad clip', 'replace a scene/segment', or 'bad take' are missing, so it matches anchor 4 rather than anchor 5's comprehensive synonym coverage.	4 / 5
Distinctiveness Conflict Risk	A clear niche with distinct triggers: 'single-take UGC master', 'Seedance', named scripts, the openai-proxy path, and billing semantics. It would not plausibly trigger any other skill, matching anchor 5's minimal conflict risk.	5 / 5
	Total	16 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: gooseworks-ai/goose-skills
Path: skills/ads/capabilities/ugc-fixloop/SKILL.md
Commit: cc3e518

Reviewed: 7 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.