fieldreport

star 0

Generate a Markdown technical report in academic paper structure for a completed task. Use this skill whenever the user asks for a 技术报告, 技术总结, 方案汇报, research report, or post-task documentation — even if they don't explicitly name the skill. Accepts an optional task page path or slug to identify the task. Sources material from `/media/yhr/2T/Canon/tasks/<task>.md` (primary) and `.planning/conversations/` / `.agent-state/conversations/` (historical runtime buffers, secondary) by default. Do not use while the task is still actively being implemented unless the user explicitly wants a report artifact at that moment.

yuanhaorannnnnn By yuanhaorannnnnn schedule Updated 6/9/2026

name: FieldReport description: | Generate a Markdown technical report in academic paper structure for a completed task. Use this skill whenever the user asks for a 技术报告, 技术总结, 方案汇报, research report, or post-task documentation — even if they don't explicitly name the skill. Accepts an optional task page path or slug to identify the task. Sources material from /media/yhr/2T/Canon/tasks/<task>.md (primary) and .planning/conversations/ / .agent-state/conversations/ (historical runtime buffers, secondary) by default. Do not use while the task is still actively being implemented unless the user explicitly wants a report artifact at that moment.

Report

Generate a 6-section technical report in academic paper structure, built from Canon task pages (/media/yhr/2T/Canon/tasks/<task>.md) as the primary source, with .planning/conversations/ and .agent-state/conversations/ as secondary historical evidence.

Canon 输出边界

读取共享契约:/home/yhr/.agents/repos/agent-skills/references/canon-output-contract.md

  • .planning/.agent-state/ 是报告证据源,不是报告后的长期 source of truth。
  • 技术报告本身是 artifact;默认保存到用户指定路径或当前 repo 合适目录,并在 Canon artifacts/ 或 update-card 中引用绝对路径。
  • 已完成任务的长期结论应更新 Canon task/project/decision/pattern/incident 页面。
  • 若证据不足只生成临时报表,最终回复要说明未做 Canon promotion。

Structure

  1. Abstract — background, problem, approach preview, key outcome
  2. Related Works — alternatives considered, why rejected
  3. Method — final technical selection, rationale, constraints and tradeoffs
  4. Implementation — code architecture, module organization, data flows
  5. Evaluation — how it was tested, results, benchmarks, vs acceptance criteria
  6. Conclusion and Future Work — summary, remaining risks, next steps

Rules

  • Build from repository evidence, not memory. If evidence is insufficient, say so.
  • Keep it narrow — about the task, not the whole repository.
  • Every claim traces back to a planning doc, commit, test result, or decision record.
  • Do not use while the task is still being implemented unless explicitly asked.

Gate

报告完成后跑:

python3 ~/.claude/skills/FieldReport/scripts/source_gate.py <report.md>

blocked → 缺节或缺源。3 项:文件存在且 ≥200 字、6 节完整、引用了至少一个 Canon task / planning doc / commit。pass → 继续 Canon promotion。

Install via CLI
npx skills add https://github.com/yuanhaorannnnnn/agent-skills --skill fieldreport
Repository Details
star Stars 0
call_split Forks 0
navigation Branch main
article Path SKILL.md
More from Creator
yuanhaorannnnnn
yuanhaorannnnnn Explore all skills →