AgentSkillsCN

zeno

为新功能生成产品需求文档(PRD)。适用于功能规划、新项目启动,或被要求撰写 PRD 时。触发关键词包括:创建 PRD、为……撰写 PRD、规划此功能、明确需求、细化规格。

SKILL.md
--- frontmatter
name: zeno
description: Persistent structured memory + telemetry + audit trail for Claude Code sessions. Use when the user wants automated capture, durable context, and standardized turn-by-turn logging.
allowed-tools: Read, Write, Bash(python:*), Bash(jq:*)

Zeno (Lean Skill)

Zeno is an evidence-first, read-only workflow for huge corpora. Keep the corpus external (JSONL REPL), retrieve only minimal slices, and cite every non-trivial claim.

How this skill is triggered (and when to use it)

Zeno activates when the user explicitly asks for Zeno or names a mode. Common triggers:

  • "Use Zeno codebase-archaeology to trace <symbol> across the repo."
  • "Run Zeno security-audit on src/ and build evidence chains."
  • "Use Zeno architecture-mapping to document entrypoints and lifecycle."
  • "Zeno pr-review these changed files: ... with downstream impact."
  • "Zeno skill-generation on this repo and draft SKILL.md."
  • "Zeno deep-research on this project question: ..."

Full use cases and mode details are defined in references/modes.md.

Hooks and automation (Claude Code)

Hooks provide always-on persistence and context injection. If the user expects standardized memory every turn, ensure the hooks are installed and verified. See:

  • references/skill_full.md (full manual)
  • references/skill_blocks.md (mandatory output blocks)

Non-negotiables

  • READ-ONLY: do not modify files unless explicitly requested later.
  • No evidence, no claim: every claim must cite path:Lx-Ly or a grep hit.
  • Small bites: prefer peek + targeted read_file slices.
  • Be deterministic: if unsure, retrieve more evidence.

Tools available

  • list_files, peek, read_file, grep, extract_symbols, stat

Budgets (summary)

  • Retrieval ops: <=30 per answer, <=12 per section
  • read_file: <=400 lines per call, <=2,000 total
  • grep: <=200 hits per call (narrow if truncated)
  • Recursion: <=12 capsules, depth <=2
  • If a budget is hit: stop, summarize, provide Next Retrieval Plan

Mode selection (quick)

  • codebase-archaeology, security-audit, architecture-mapping, pr-review, skill-generation, deep-research Detailed mode playbooks: references/modes.md.

Mandatory output blocks

Always append the three machine-parseable blocks at the end of every response. Format is defined in:

  • references/skill_blocks.md

NON-NEGOTIABLE: Read the full manual

Before using Zeno in a real task, you MUST read the full manual and follow it. This lean SKILL is only a pointer.

Read these in order:

  • references/skill_full.md (complete operating manual, required)
  • references/skill_blocks.md (mandatory output blocks, required)
  • references/operator_instructions_and_review.md (operator-grade rules)
  • references/recipes.md (copy/paste retrieval plans)
  • references/protocol.md (JSONL REPL protocol)
  • references/security.md and references/otel.md