check-mailbox
Skill8 starsChanged 11 days agoApache-2.0View on GitHub
Check your agent mailbox for messages at session start. Use at the beginning of every session to review communications from other agents or advisors and note any action items.
In this repository39 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- .cursor/rules/architecture-patterns.mdcCursor rule
- .cursor/rules/completion-discipline.mdCursor rule
- .cursor/rules/github-tracking.mdcCursor rule
- .cursor/rules/programmer-briefing.mdcCursor rule
- .cursor/rules/verification-first.mdcCursor rule
- ab-a-isolationSkill
- assign-sprint-safelySkill
- audit-cascadeSkill
- brief-coding-agentSkill
- cleanup-dev-activeSkill
- close-issue-properlySkill
- cohort-attention-rollupSkill
- compost-reviewSkill
- continue-narrativeSkill
- create-omnibusSkill
- create-session-logSkill
- cut-releaseSkill
- delete-module-safelySkill
- deliver-mailSkill
- discovered-work-captureSkill
- doc-sync-sweepSkill
- draft-blog-postSkill
- draft-weekly-shipSkill
- duty-cycle-tickSkill
- narrative-verificationSkill
- piper-draft-issueSkill
- piper-draft-specSkill
- piper-sprint-planSkill
- piper-stakeholder-updateSkill
- piper-synthesize-feedbackSkill
- propose-featureSkill
- publish-to-blogSkill
- query-github-boardSkill
- template-auditSkill
- trust-checkSkill
- update-calendarSkill
- update-current-stateSkill
- update-piperSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Check your agent mailbox for messages at session start and process any pending communications.
- Distinctive: line 18
When PM mentions you have messages
- Distinctive: line 19
When resuming after a break
- Distinctive: line 20
After another agent references sending you something
- Distinctive: line 47
Read the full message
- Distinctive: line 56
After reading each message:
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
check-mailbox
Check your agent mailbox for messages at session start and process any pending communications.
When to Use
Use this skill:
- At the start of every session (before beginning assigned work)
- When PM mentions you have messages
- When resuming after a break
- After another agent references sending you something
Procedure
Step 1: Check Your Inbox
ls mailboxes/{your-role-slug}/inbox/Role slugs: arch, cio, lead, comms, ppm, cxo, host, docs, exec, spec
Mail v4 too, if your role has joined it (every seat in this repo from 2026-11-09; xian's yes 2026-11-05). If your role is pilot: true in ~/Development/piper-morgan-mail/roles.yaml, also run:
python3 scripts/mail4.py inbox --role {your-role-slug}Read each message with scripts/mail4.py read ID, act on it, then scripts/mail4.py ack ID done (or read / declined). Your first inbox of the day also runs the daily canary. Mail to in-repo seats goes through scripts/mail4.py send; v3 (mail-send.sh) stays as the fallback until 2026-11-23. The mailbox is empty only when both are. Ops: docs/internal/operations/mail-v4-pilot.md.
If empty: Note "Mailbox empty" in session log and proceed with assigned work.
If messages exist: Continue to Step 2.
Step 2: Read Each Message
For each file in inbox:
- Read the full message
- Note the headers:
From: Who sent itResponse-Requested: yes/noPriority: high/medium/low
- Understand what's being asked or communicated
Step 3: Process and Move
After reading each message:
mv mailboxes/{your-slug}/inbox/{filename} mailboxes/{your-slug}/read/Why move? Messages in inbox/ = unread. Moving to read/ prevents re-reading next session.
Move only what you read, by name. Copy each filename from your own read or skim output into the mv. Never derive the list from the inbox: a glob, for f in $(ls inbox) and ls | xargs mv all file whatever synced since you looked, unread.
The mv alone is not durable. It changes only your worktree. To land the move on origin/main, send it through scripts/mail-send.sh in one call, passing BOTH sides of the move (the inbox/ path and the read/ path) plus any outbound memo paths: scripts/mail-send.sh "mail({role}): triage N" mailboxes/{you}/inbox/{f} mailboxes/{you}/read/{f}. Never git add/git commit a mailbox path. Do not cc PM or address mail to PM (route PM-needed items to exec). Full workflow: CLAUDE.md, "The mailbox workflow" (refreshed 2026-10-05 by Docs, #1909 skills review).
Step 4: Handle Response-Requested
If Response-Requested: yes:
- Create response in sender's inbox:
mailboxes/{sender-slug}/inbox/{your-response-filename}.md - Use response format:
# Re: [Original Subject] **From**: {your-slug} **To**: {sender-slug} **Date**: YYYY-MM-DD HH:MM **In-Reply-To**: {original-filename} **Response-Requested**: no **Priority**: medium --- [Your response] - If you can't respond yet: Note in session log as pending action item.
Step 5: Note Action Items
In your session log, record:
- Messages received (count and from whom)
- Key decisions or information
- Any action items for this session
- Pending responses needed
A ruling of xian's is recorded and relayed the same hour. When you hear one (in conversation, or relayed in a memo), add its line to docs/internal/architecture/decisions/decisions.log and mail every seat it moves within the hour, not at wrap-up. Before you idle, check that each relay landed: the memo is in their inbox on origin/main (git ls-tree origin/main mailboxes/{them}/inbox/).
Anti-Patterns to Avoid
| Don't Do This | Why | Do This Instead |
| Skip mailbox check | Miss important context or assignments | Always check at session start |
| Leave messages in inbox | Re-read same messages next session | Move to read/ after processing |
| Ignore Response-Requested | Sender waiting indefinitely | Respond or note as pending |
| Read but don't log | No record of what you learned | Note key points in session log |
Example: Full Mailbox Check
# 1. Check inbox
ls mailboxes/docs/inbox/
# Output: memo-cio-skill-response-2026-01-21.md
# 2. Read message
# [Use Read tool on the file]
# 3. Move to read/
mv mailboxes/docs/inbox/memo-cio-skill-response-2026-01-21.md mailboxes/docs/read/
# 4. If Response-Requested: yes, create response in sender's inbox
# [Create response file in mailboxes/cio/inbox/]
# 5. Log in session log:
# "### 8:00 AM - Mailbox Check
# - Received 1 message from CIO re: skill adoption
# - Decision: Approved with guidance
# - Action: Proceed with Tier 1 skills"Example: Empty Mailbox
# 1. Check inbox
ls mailboxes/lead/inbox/
# Output: (empty)
# 2. Log in session log:
# "### 9:00 AM - Mailbox Check
# - Mailbox empty, no pending messages"Quality Checklist
After checking mailbox:
- Listed inbox contents
- Read all messages
- Moved all to
read/folder - Created responses where
Response-Requested: yes - Noted messages in session log
- Identified action items for session
Reference
- Mailbox system docs:
mailboxes/DIRECTORY.md - Message format: See README for full header spec
- Role slugs:
arch,cio,ceo,lead,comms,ppm,cxo,host,exec,spec,docs
Changelog
- v1.3 (2026-11-06, CIO): Step 3 moves only the files you read, by name, never a list taken from the inbox. CIO did the latter on 11-04 and 11-05 (no unread mail lost either time); Agent 360 v0.6 answer 5.4.
- v1.2 (2026-11-05, CIO): Step 1 also reads mail v4 for a role that has joined it; every seat joins 2026-11-09 on xian's yes (decisions.log 2026-11-05 10:47).
- v1.1 (2026-11-04, CIO): Step 5 records a ruling of xian's the same hour and checks a relay landed. #1715's FOLD candidate (
docs/internal/operations/skill-candidates-review.md), xian's "Yes same-hour in check-mailbox" relayed by Exec 11-04.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

