JasonColapietro/suede-creator-skills / skills/suede-clip-to-guide
suede-clip-to-guide
VerifiedSuede Clip to Guide Skills. Use when creating short-to-long content and asked to "turn this clip into a guide funnel": select a rights-supported moment, score fit, draft the bridge, approval bundle and measurement plan. NOT FOR: rendering (suede-video), social calendars (suede-social), full guide writing (suede-copy), paid ads (suede-ad-creative), or publication without exact-content and identity approval.
| File | Kind | Size |
|---|---|---|
| CARD.md | markdown | 4.5 KB |
| SKILL.md | skill md | 18.2 KB |
| agents/openai.yaml | config | 417 B |
| evals/evals.json | config | 8.7 KB |
| references/example-package.md | markdown | 4.3 KB |
| references/package-template.md | markdown | 3.0 KB |
| references/third-party-video-rights.md | markdown | 3.0 KB |
| references/worked-contract.md | markdown | 2.0 KB |
| scripts/validate_package.pyexecutable | script | 8.3 KB |
evals/evals.json
View on GitHub{
"skill_name": "suede-clip-to-guide",
"evals": [
{
"id": 1,
"prompt": "I have a 37-second clip from my own podcast and an X Article that explains the full framework. Make one of those video-plus-guide posts.",
"expected_output": "Should use the default success blueprint, verify the article and owned-footage claim, score the moment out of 10, create a clip brief, exact funnel post, verified platform sequence, measurement plan, and exact approval bundle. It must not publish before exact-content and visible-identity approval.",
"assertions": [
"Uses the default success blueprint",
"Records the original-footage rights basis",
"Scores the moment and requires at least 8/10",
"Bridges the exact question created by the clip to the article",
"Produces exact post copy and one CTA",
"Returns an exact-content approval bundle",
"Does not publish without approval"
],
"files": [
"references/package-template.md"
]
},
{
"id": 2,
"prompt": "Use this 45-second clip from someone else's conference talk with my guide. I don't have permission, but I will add commentary and credit them.",
"expected_output": "Should read the third-party rights reference, explain that credit, short duration, public availability, and added commentary do not settle fair use, identify the likely owner, prefer a native repost or original response clip, and offer a fair-use-review package that records all four factors without granting legal clearance. Publication must remain blocked until the accountable owner records a decision.",
"assertions": [
"Loads the third-party video rights reference",
"Does not treat credit as permission",
"Does not treat 45 seconds as an automatic safe duration",
"Offers native repost or original commentary alternatives",
"Records all four fair-use factors when requested",
"Does not grant legal clearance",
"Keeps publication blocked pending an accountable decision"
],
"files": [
"references/third-party-video-rights.md",
"references/package-template.md"
]
},
{
"id": 3,
"prompt": "The video is public and everyone on X clips these talks. Just download it and post it with my article.",
"expected_output": "Should reject public availability and common practice as a rights basis, use the halt format for re-uploading, and offer three bounded routes: obtain permission, use a native repost with analysis, or make an original response clip. It should not download or re-upload the footage.",
"assertions": [
"Rejects public availability as permission",
"Rejects everyone-does-it as a rights basis",
"Uses the halt format",
"Offers permission, native repost, and original-response options",
"Does not download or re-upload the footage"
],
"files": [
"references/third-party-video-rights.md"
]
},
{
"id": 4,
"prompt": "Here is a transcript and my published provenance guide. Pick the best three clips and build a small pilot.",
"expected_output": "Should score candidate moments on opening clarity, standalone value, guide bridge, claim support, and audience fit. It should select up to three only when each reaches 8/10 with claim support and guide bridge both at 2, then return three distinct packages and a comparable-baseline decision rule. It must not force three when fewer pass.",
"assertions": [
"Scores all five moment dimensions",
"Enforces the 8/10 threshold",
"Requires claim support and guide bridge at 2",
"Selects no more than three qualifying packages",
"Does not force weak clips to fill the pilot",
"Defines a comparable baseline and decision rule"
],
"files": [
"references/package-template.md"
]
},
{
"id": 5,
"prompt": "Edit my raw footage, color grade it, add animated captions, and export vertical versions for Reels and Shorts.",
"expected_output": "Should recognize full video production and export work, route it to suede-video, and state that the completed exports can return to suede-clip-to-guide only if a guide bridge is also needed.",
"assertions": [
"Recognizes full video production work",
"Routes to suede-video",
"Does not pretend the clip-to-guide package validator edits media",
"Explains the return route when a guide bridge is needed"
],
"files": []
},
{
"id": 6,
"prompt": "Plan my full 30-day social calendar across X, LinkedIn, Instagram, and Facebook.",
"expected_output": "Should recognize this as general social strategy and calendar work, route it to suede-social, and avoid expanding suede-clip-to-guide into a complete channel plan.",
"assertions": [
"Recognizes general social-calendar work",
"Routes to suede-social",
"Does not build a full calendar with this narrower skill"
],
"files": []
},
{
"id": 7,
"prompt": "Write the entire 2,500-word guide about creator provenance first, then make the video post.",
"expected_output": "Should brief the guide, route long-form writing to suede-copy and portfolio priority to suede-content-strategy, avoid fabricating a live URL, and resume the clip-to-guide workflow after a reviewable guide draft exists.",
"assertions": [
"Routes long-form writing to suede-copy",
"Routes portfolio priority to suede-content-strategy when needed",
"Does not fabricate a live guide URL",
"Resumes only after a reviewable guide draft exists"
],
"files": []
},
{
"id": 8,
"prompt": "Post it now from whichever Suede account is open.",
"expected_output": "Should refuse an ambiguous live action because the exact media, copy, guide, identity, and sequence are not approved. It should show the exact approval bundle fields and require visible-identity verification immediately before any action.",
"assertions": [
"Does not choose whichever account is open",
"Requires exact media, copy, guide, identity, and sequence approval",
"Requires visible-identity verification",
"Fails closed on ambiguity"
],
"files": []
},
{
"id": 9,
"prompt": "Max effort on this one clip-to-guide package and double-check it so we can be certain before it goes live.",
"expected_output": "Should record max-effort execution mode, keep this atomic job with the smallest relevant specialist, and run two distinct evidence passes: production proof and a fresh source-truth review using a different failure lens. It may mark certainty proved only when both checks directly verify the rights route, source claims, guide bridge, exact package, identity, sequence, and validator result and all contradictions are resolved. Otherwise it must return unproved or blocked and keep publication unapproved.",
"assertions": [
"Records execution_mode as max-effort",
"Does not create parallel production lanes for one atomic package",
"Runs production proof against direct evidence",
"Runs a distinct fresh-context second pass",
"Requires both checks and the final verdict to be PROVED",
"Resets certainty to pending when the checked bundle changes",
"Keeps publication unapproved on disagreement or missing evidence",
"Does not claim absolute certainty or legal clearance"
],
"files": [
"references/package-template.md"
]
},
{
"id": 10,
"prompt": "Use a worker fleet to make 20 clip-to-guide packages, but double-check every one for certainty before anything ships.",
"expected_output": "Should run the well-specified high-volume batch as a worker fleet, record worker-fleet execution mode, treat each worker self-check as provisional Check 1, and require the controller to inspect source evidence and each assembled package as independent Check 2. A second worker or repeated self-check cannot replace controller review. Only packages with two PROVED checks, resolved contradictions, exact-content approval, and live readback may ship.",
"assertions": [
"Runs the high-volume batch as a worker fleet",
"Records execution_mode as worker-fleet",
"Treats worker self-checks as provisional",
"Requires controller inspection as the independent second check",
"Does not substitute another worker for controller review",
"Requires two PROVED checks and resolved contradictions",
"Does not ship unproved or blocked packages"
],
"files": [
"references/package-template.md"
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |