Conference Foresight Packaging

What this skill is for

Use this skill to package ACP's Foresight-facing evidence, framing, limitations, artifact index, and demo script.

When to use it

• Preparing artifacts/conference/foresight.
• Reviewing whether the packet is submission pre-packaging ready.
• Checking that contribution list and limitations are present.

When not to use it

• Do not use it to make final camera-ready prose.
• Do not use it to inflate benchmark evidence into field claims.
• Do not use it before the stabilization boundary is understood.

Inputs expected

• Target output directory
• Foresight venue frame
• Benchmark and rehearsal artifacts
• Limitations/non-claims

Preflight / prerequisites

1. Confirm stabilization packet status.
2. Confirm benchmark, ablation, and scale-band commands pass.
3. Confirm the package includes limitations and artifact hashes.

Workflow

1. Run npm run conference:foresight.
2. Inspect package summary and artifact index.
3. Confirm reviewer start, abstract, contribution list, casebook, figure plan, limitations, and demo script exist.
4. Run release gate before claiming package readiness.

Decision rules / judgment criteria

• Reviewer legibility matters more than artifact volume.
• Every claim should point to a generated or source artifact.
• Limitations must be adjacent to evidence.

Escalation rules

Escalate when artifact index is missing, required memos are absent, gates fail, or the requested claim is not supported by artifacts.

Available scripts

• scripts/smoke.mjs: checks expected Foresight package files.

Outputs

• Foresight package readiness notes
• Missing artifact list
• Gate recommendation

Failure handling

Do not delete failed package output; preserve and report what is missing.

Trust / safety notes

Conference packaging is claim-bearing. Keep prototype, benchmark, and field-efficacy boundaries explicit.

Composition notes

Use in foresight-submission-prep after public-hearing evidence generation.

Examples to inspect next

Read examples/casebook.md for packaging success, missing limitations, ambiguous evidence, and overclaiming.

Evaluation hooks

Run scripts/smoke.mjs, npm run conference:foresight, and npm run skills:audit.