writing-workbench/projects/水晶裹尸布与生命体/editor/韩愈/r1.1/agent-invocation-packet.md

4.4 KiB
Raw Blame History

Agent Invocation Packet: han-yu

1. Invocation Identity

  • project_id: 水晶裹尸布与生命体
  • workflow: article-outline-development
  • stage: outline-repair
  • run_id: r1.1
  • participant_id: han-yu
  • target_participant_name: 韩愈
  • participant_type: ccpe_lite
  • canonical_artifact_path: C:\Users\wangq\Documents\Codex\ccpe-system\agents\lite\han-yu.prompt.md
  • canonical_artifact_version: current
  • invocation_mode: prompt_path_reference
  • packet_type: participant_startup_packet

2. Carrier Metadata

  • carrier_type: agent_subsession_participant
  • thread_id: not_applicable_platform_subagent
  • agent_id: 019ee37a-85e9-7d11-a2cf-bc7e7a496095
  • persistent_required: true
  • resume_or_followup_method: multi_agent_v1.send_input with recorded agent_id
  • execution_status: returned
  • strict_invocation_required: true
  • simulation_only: false
  • returned_output_path: C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\editor\韩愈\r1.1\returned-output.md
  • skill_execution_record_path: not_applicable
  • prompt_to_send_path: C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\editor\韩愈\r1.1\prompt-to-send.md

3. Canonical Artifact Handling

  • should_paste_full_prompt_or_spec: no
  • role_integrity_requirement: must_load_canonical_artifact_before_task
  • allowed_wrapper_instructions: project_id, workflow, stage, run_id, input_files, context_files, output_path, output_language_policy
  • forbidden_role_changes: replacing the canonical role; replacing the canonical tone; replacing the canonical output structure; weakening the canonical writing method; treating the task prompt as the role definition
  • no_simulation_requirement: true
  • path_readability_evidence: controller preflight read succeeded on 2026-06-20T13:20:38+08:00

4. Current Task

汇总 r1.1 三份评审报告的意见判断其合理性然后输出《水晶裹尸布与生命体》v4.0 版提纲。

This run is a first test invocation of 韩愈 as lead writer. The user explicitly authorizes direct progression for this test: do not stop to ask the user to confirm the structure archetype before producing v4.0. Still disclose the selected structure archetype and why.

5. Input Files

  • C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\outline\水晶裹尸布与生命体提纲3.0.md

6. Context Files

  • C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\intake\背景资料.md
  • C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\audit\张辽\r1.1\returned-output.md
  • C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\audit\显影\r1.1\returned-output.md
  • C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\audit\巨人\r1.1\returned-output.md

7. Embedded Inputs

No embedded content. Use the local file paths above.

8. Output Contract

  • source_primary_language: Chinese
  • output_language_policy: mirror_source_primary_language
  • human_review_language_requirement: Chinese
  • human_confirmation_artifact: false
  • gate_id_if_applicable: not_applicable
  • expected_output_type: review_judgment_plus_outline_v4

Required output shape:

  1. 简短说明实际读取的输入。
  2. 简短评审意见取舍判断:采纳 / 部分采纳 / 不采纳,并说明理由。
  3. 输出完整 v4.0 版提纲。

9. Continuity Policy

This is 韩愈's first editor session and first round for this project. Preserve the same agent_id for future follow-up rounds such as r1.2 if the user asks 韩愈 to continue.

10. Session Logging

The controller records this packet, prompt-to-send.md, run-metadata.json, agent_id, and returned-output.md under this run directory.

11. Forbidden Actions

  • Do not simulate reviewer voices.
  • Do not write a main-session synthesis outside returned-output.md.
  • Do not generate article正文.
  • Do not create or modify outline files; the user will extract formal content manually.
  • Do not overwrite files outside this run directory.

12. Return Path

Write the formal returned output to:

C:\Users\wangq\Documents\Codex\writing-workbench\projects\水晶裹尸布与生命体\editor\韩愈\r1.1\returned-output.md

13. Runtime Acceptance Rule

The output is formal only if a real child agent loaded or read the canonical artifact before task context, returned a result, and returned-output.md records simulation_only: false.