القائمة
Use only when autopilot-draft dispatches document strategy and evidence-plan creation. Not for top-level user requests or primary capability routing.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
استنادا إلى تصنيف SOC المهني
| name | draft-strategy |
| description | Use only when autopilot-draft dispatches document strategy and evidence-plan creation. Not for top-level user requests or primary capability routing. |
| argument-hint | <mode> --inputs <comma-separated-paths> --output <artifact-dir> [--intensity direct|quick|standard|strong|thorough|adversarial] <task description> |
| metadata | {"group":"sub","fam":"sub","invocation_class":"parent-invoked","modes":["rebuttal","paper","review","report","proposal","presentation"],"blurb":"Create an initial document strategy and evidence-based writing plan.","use_when":"Use only when autopilot-draft dispatches document strategy and evidence-plan creation.","not_for":"Not for top-level user requests or primary capability routing."} |
Follow an explicit target, venue, audience, or source-artifact language first. Otherwise, write strategy artifacts and user-facing output in the conversation language according to <agent-home>/roles/response-policy.md. Preserve quotations, code identifiers, citations, paths, and venue-specific terms when translation would reduce precision.
Parse $ARGUMENTS as follows:
mode: rebuttal | paper | review | report | proposal | presentation--inputs <comma-separated-paths>: pre-discovered artifact directories, usually analysis_project/{paper,doc}/... or research/{topic}/--output <dir>: <artifact-root>/documents/{date}_{name}/--intensity: derive the quick | light | standard | thorough | adversarial verification tier through CONVENTIONS §1.1Verify the expected analysis files under {output_dir}/analysis/:
material_index.mdreviewer_analysis.mdref_analysis.mdIf a required file is missing, report that autopilot-draft Step 1 did not complete; do not invent the analysis.
Select one strategy template from the first argument:
| Mode | Strategy focus |
|---|---|
rebuttal | Meta-review, priority matrix, and reviewer-by-reviewer response |
paper | Positioning, contribution, outline, and evidence |
review | Evidence-grounded peer-review response |
report | Objective, findings, and section plan |
proposal | Problem, approach, work plan, and impact |
presentation | Audience, core message, and slide outline |
autopilot-draft maps its form-first paper, presentation, or doc route to these six labels during preflight. A doc request maps to rebuttal, review, report, or proposal from task intent, with report as the final fallback. A direct draft-strategy invocation must supply one of the six labels.
Read references/delegate-prompt.md for the complete mode mapping, six templates, paragraph-cohesion precheck, tone detection, slide conventions, and quality requirements.
{output_dir}/analysis/.references/delegate-prompt.md and dispatch the research/research-survey unit. Require it to write the strategy file directly and return only paths plus a 3–5 line summary.references/qa-review.md; run the selected reviewers in parallel and allow at most two rounds.editorial/translate unit with references/mirror.md; do not infer a fixed mirror language from the conversation.| File | Load when | Contents |
|---|---|---|
references/delegate-prompt.md | Building the research/research-survey unit prompt | Inputs, paragraph-cohesion precheck, mode mapping, six strategy templates, tone detection, slide conventions, quality requirements, and return contract |
references/qa-review.md | Selecting and running QA | Rigor scaling, fast fact-check rationale, reviewer and fact-checker prompts, verdict branches, and two-round cap |
references/mirror.md | An explicit companion-language artifact is required | editorial/translate unit procedure, primary-language decision, and final report line |
$ARGUMENTS
Use when source code must be implemented, debugged, refactored, or code-audited through a plan, execution, test, and report loop. Not for requirements-only changes, new empirical experiments, or document-only editing.
Use when persistent analysis of code, a paper, or a document must be created or refreshed because it is absent, stale, or explicitly requested. Not for read-only project orientation, context recovery, or status reporting.
Use when durable cross-project user preferences must be inferred from coding, writing, or analysis evidence and stored as a profile. Not for one-project context recovery, casual preference acknowledgment, or ordinary task execution.
Use when a visual product surface needs references, design tokens, components or mockups, review, and development handoff. Not for implementing an already-approved design in code or for document prose work.
Use when a new paper, presentation, report, proposal, or other user-facing document must be produced from evidence. Not for correcting only an existing document or for source-code implementation.
Use only when autopilot-code dispatches the implementation stage for an approved plan. Not for top-level user requests or primary capability routing.