Suede Clip to Guide

Suede Clip to Guide

Build one evidence-backed path from a video moment to a deeper written asset. Treat the clip as the attention layer and the guide as the authority layer. Do not assume the pattern performs; test the bridge against the account baseline.

Default Success Blueprint

Use this as the recommended pattern when the rights, source, claim, platform, and approval gates pass:

Useful guide with one promise
        ↓
Self-contained video moment
        ↓
Rights route + 8/10 moment score
        ↓
Clip hook + source credit + subtitles
        ↓
Post explains why the moment matters
        ↓
One explicit bridge to the guide
        ↓
Max Effort/Fleet: dual certainty gate
        ↓
Exact-content approval
        ↓
Public readback + clip-to-guide measurement
        ↓
Keep, revise, or stop from the observed result

Prefer this blueprint when:

• a guide already answers the natural next question created by the clip;
• the goal is deeper understanding, bookmarks, qualified replies, leads, or authority rather than raw views alone;
• the selected moment passes the 8/10 score and claim-support threshold;
• the rights route and current platform sequence are recorded;
• the account can observe at least one clip metric and one guide-transition or campaign metric.

Do not use it when:

• the guide is thin, unrelated, unpublished without a stable draft, or has no specific promise;
• the clip needs missing context, distorts the source, or exists only because it is popular;
• third-party reuse is blocked or the accountable fair-use decision is absent;
• the platform cannot preserve the source, clip, and guide bridge in a verified sequence;
• the idea is complete in the short post and the guide would add no depth.

For a first test, build three distinct packages only when three moments pass all gates. Compare them with the account's recent comparable posts. Keep the pattern only if the named guide-transition or campaign metric meets the decision rule; revise one variable at a time when it does not.

Non-Negotiable Gates

Apply these throughout the run. Resolve rights, source, and claims before writing copy; resolve platform, certainty, and approval before a live action:

1. Rights gate: classify the video as owned, licensed, permission-recorded, native-repost-only, fair-use-review, or blocked. Public availability is not reuse permission.
2. Source gate: obtain the video, transcript or exact timestamps, guide draft or URL, target platform, intended account, and desired next action.
3. Claim gate: trace factual claims and quotations to the source. Label interpretations as interpretations.
4. Platform gate: inspect current official requirements and the live composer before promising that media, quote-posts, article cards, links, or replies can be combined.
5. Approval gate: do not publish, schedule, repost, upload, reply, or edit a live article until the user approves the exact media, copy, guide, account, and sequence.
6. Certainty gate: when a max-effort or worker-fleet controller runs this skill, require two distinct evidence checks before marking a package approved or published. A worker conclusion is not the second check.

When a blocking gate fails, stop the affected action and return:

HALT, <one-line blocker>
Options:
1. <bounded resolution>
2. <bounded resolution>
3. <bounded resolution, if useful>
Waiting on: <specific evidence or approval>

Continue only on independent draft work that does not bypass the blocker.

Rationalizations that precede a violation

Each excuse below shows up immediately before a gate gets skipped. When one appears in your own reasoning or in the request, treat it as the signal to run the gate, not to waive it.

Workflow

1. Define the funnel contract

Capture:

• campaign goal: article reads, bookmarks, qualified replies, profile visits, leads, or another named action;
• target platform and visible posting identity;
• video source URL or file, owner, transcript availability, and candidate timestamps;
• guide title, URL or draft, its promise, and current publication state;
• audience and one next action;
• whether the user wants a private draft, scheduled package, or live execution.

If product-marketing context exists in .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md, read it before asking for information already covered there.

2. Choose a lawful video route

For any source that is not original, licensed, or permission-recorded, read references/third-party-video-rights.md. Do not infer fair use, ownership, or permission. Shortness, credit, a public source, or added commentary alone does not settle fair use. Do not remove watermarks, attribution, or provenance. Prefer a native repost with added analysis or create an original camera or screen-recorded response that summarizes the idea without copying the footage.

3. Establish the guide anchor

The long-form asset must have:

• one specific promise that the clip can honestly introduce;
• a title and stable draft or URL;
• enough depth to reward the transition from short to long;
• sources for material factual claims;
• one next action that matches the campaign goal.

If the guide does not exist, return a brief and route the full draft to suede-copy and the portfolio decision to suede-content-strategy. Resume this skill after the guide has a reviewable draft. Do not fabricate a live URL.

4. Score candidate moments

Score each candidate from 0 to 2 on five dimensions:

Select a moment only when it scores at least 8/10, claim support is 2, guide bridge is 2, and the rights gate passes. If none qualify, return the scorecard and request a better source instead of forcing a clip.

5. Build the package

Read references/package-template.md and produce:

1. Decision snapshot: goal, audience, platform, identity, and chosen sequence.
2. Source and rights record: owner, source, evidence, transcript excerpt, permitted route, jurisdiction, and any four-factor review.
3. Guide anchor: exact title, promise, URL or draft state, and the question it answers after the clip.
4. Clip brief: in/out timestamps, target duration, exact opening, subtitle text, on-screen credit, ending bridge, and export owner.
5. Funnel post: exact text with four jobs:
   • hook the video moment;
   • explain why it matters;
   • bridge to the deeper guide;
   • give one specific next action.
6. Publish sequence: one of the verified patterns below.
7. Measurement plan: baseline, observation window, and decision rule.
8. Certainty record: execution mode, two check owners or processes, direct evidence, contradictions, and final verdict.
9. Approval bundle: exact media, post copy, guide, identity, and sequence.

Do not pad the post with generic setup, unsupported performance language, or multiple competing calls to action.