human-ai-schedule

name	human-ai-schedule
description	Co-author doctrine for shared, half-hour-block daily schedules. Make sure to consult this skill whenever working in a dated daily log, journal, retro, or coordination file shared with a coworker — even if the user only says "log this at 11:00" or "what did we say last Tuesday." Default behavior of dumping content at the bottom of files or paraphrasing the user's voice into corporate prose must defer to this skill, which enforces four jobs: locate, format, commit, harness. TRIGGER when: user is in a file matching `daily/`, `journal/`, `schedule/`, `半小时复盘/`, or any file with H1 time-block headers like `# 7：30` or `# 09:00`; user says "create today's", "创建 今天的", "log at N o'clock", "补到 N 点那格", "morning brief", "晨会复盘", "what did we decide on X day"; user dictates an event with time anchor ("just now", "刚才", "yesterday", "上周三", explicit clock); user asks to reconstruct a past session; user references a tag aggregation page; agent receives a `#协作` / `#collab` segment and needs to leave a commit. SKIP when: pure code editing without log context; one-off chat with no persistent dated file; user explicitly says "just answer, don't log"; reading code or technical docs only. Examples that trigger: "创建今天的 schedule", "log this at 11:00", "把刚才那段补到 11 点那格", "what did we say last Tuesday", "翻 5.7 那次 session", "记一下今天去了彭镇喝茶". Examples that skip: "implement a stripe webhook", "explain async", "改一下 settings.json 的 permission", "rename this private function".
extends	ctrl-c-v

Human-AI Schedule

You are the same engineer. Same Friday flight to the Alps. Same system. But you are not alone in the office anymore — there is a coworker, and chat history forgets what was said by Wednesday. So you built a third thing: a shared schedule, half-hour-block, dated, tagged, persistent. Both write to it. Either can leave a commit on the other's entries.

The schedule remembers what neither of you can. Without exception.

---

§ H0 — Co-author role

Four jobs every entry:

□  Locate  — session timestamp → date + time-block decided before write
□  Format  — H1 time / H2 #tag / single `---` separator preserved
□  Commit  — `#协作` present? → § H4 evaluation → write or stay silent
□  Harness — project-tagged? → aggregation row appended same turn

Anti-patterns (revert immediately if seen):

Pattern	Why wrong
Content written without time-block anchor	floating, unrecoverable
Empty time-block filled with speculative content	noise inflation
User's prose paraphrased into corporate tone	voice lost
Timestamp invented or assumed	unverifiable, breaks forensics

§ H1 — Locate before writing

Session wall-clock timestamp is the truth source for "now." Convert to user's local timezone, then walk this tree:

User dictates content
    │
    ├─ explicit time stated → use that time → convert relatives to absolute date in prose
    ├─ contains "now / just now / 刚才 / 现在" → session timestamp's current block
    ├─ past event, no time stated → ASK user (do not write yet)
    └─ meta-discussion (workflow, tooling, retro) → session timestamp's current block

Decision tree details and forensics → @${CLAUDE_SKILL_DIR}/playbooks/locate.md

§ H2 — File skeleton (project-defined)

Skeleton (filename, granularity, top section) is project-defined in SCHEDULE_TEMPLATE.md. Universal invariants:

□  Each time-block is an H1 header
□  Tag segments under a time-block are H2, prefixed `#`
□  Separator between time-blocks is a single `---`
□  Colon style (full-width / half) consistent file-wide — match what exists
□  New day = copy template → empty blocks → no speculative entries

§ H3 — Tags (vocabulary project-defined, rules universal)

Vocabulary lives in SCHEDULE_TEMPLATE.md. Universal rules:

□  Multiple tags per block allowed: `## #exercise #leisure`
□  Project-named tags take precedence over generic
□  `#协作` / `#collab` stacks with other tags, never replaces
□  Capitalization stable per tag — never auto-normalize
□  New project tag → register in aggregation page first → then use

Aggregation mechanics → @${CLAUDE_SKILL_DIR}/playbooks/aggregate.md

§ H4 — Dual-signature commits

Trigger: #协作 / #collab tag present in the segment.

Format:

- #commit (user-handle): user's view — decision, next action, judgment
- #commit (agent-id): agent's view — risk, dependency, uncertainty

Identity: users sign with handle. Agents sign with model ID (claude-opus-4.7, gpt-5) — model versions produce different judgment, signature lets future readers calibrate.

Write a commit only when:

□  noticed risk the other side did not raise
□  hold related context they did not ask for
□  must disclose uncertainty about your own work

Do NOT write:

Pattern	Why wrong
"great job" / "looks good"	affirmation, zero info
restating the outcome	adds nothing
"I'll keep watching this"	non-commitment, noise
general encouragement	not your job

Silence is the default.

§ H5 — Voice (anti-AI-tone)

Logs are diary, not corporate prose. User's voice — casual, specific, sometimes profane — survives.

□  Bullet density low: 3 lines of prose > 5 bullets
□  Verbs present: "标签格式统一了" not "标签管理体系完善"
□  Cause / blocker / outcome shape, not flat outcome
□  Profanity preserved if user used it

Detailed rules + reference rewrites → @${CLAUDE_SKILL_DIR}/playbooks/tone.md

§ H6 — Cross-session forensics

User asks "what did we say on day X" → walk:

1. List session logs by mtime → identify candidates near day X
2. Filter by absolute date in timestamp (NEVER by file mtime)
3. Extract user + agent messages with timestamps
4. Surface to user → confirm content matches their memory
5. Write into dated file ONLY after confirmation

Forensics commands and recipes → @${CLAUDE_SKILL_DIR}/playbooks/jsonl-forensics.md

§ H7 — Tag aggregation page

Single index file (tag-index.md or equivalent) lists every project-tagged time-block as one row.

□  One time-block = one row (never compress days)
□  Link uses `file#time-anchor` syntax
□  Time format matches source H1 exactly (full-width if source full-width)
□  Append immediately when project tag added — never batch to end-of-week

Schedule (chronological) and aggregation (by topic) are two views of same data. Must agree.

§ H8 — Trigger table

User says / does	Agent does
"create today's" / "创建今天的"	verify session date → copy template → empty time-blocks
"format this md"	normalize separators, tag levels, spacing — do NOT rewrite prose
dictates current event	session timestamp's current block → write
dictates timed past event	use stated time → convert relatives to absolute dates
dictates past event, no time	ASK user (do not write yet)
"go find day X"	§ H6 forensics → confirm with user → write
"make X concise"	three-paragraph form (cause / blocker / outcome), not bullets
#协作 segment present	§ H4 evaluation — write commit only if risk/dep/uncertainty
project-tagged segment added	mirror to aggregation page same turn

---

— Completion criteria —

□  Locatable    — every entry has a date and a time-block
□  Voice kept   — six months later, user recognizes their own voice
□  Recoverable  — "what did we decide on day X" answered in 60s
                  from schedule + aggregation, no chat history needed

---

The schedule held the line. See you on the slopes. ⛷️