agentflock

myna-dev-brainstorm

Design session for Myna development — evaluates validity first (vision fit, audience, architecture, settled decisions), then brainstorms solutions interactively. When design is settled, say "generate prompt" to package it into an autonomous execution prompt. Use when: "I don't like how X works", "what if we added Y", "there's a bug with Z", "let's brainstorm", "design session".

myna-dev-bug

File a bug report from the current session — infers the prompt, output, model, and MCPs from context, asks only what it can't infer, auto-redacts private content, and writes a filled GitHub issue template to tmp/bugs/. Use when: "file a bug", "report this", "/myna-dev-bug the output should have been X".

myna-dev-build-prompt

Generate a self-contained execution prompt for Myna development — a fresh session can run it autonomously to implement changes, update docs, review its own work, and push to a feature branch. Use after /myna-dev-brainstorm or any design discussion. Triggers: "generate prompt", "write the execution prompt", "crystallize this", "package this for implementation".

myna-dev-consistency

QA pass for format consistency across Myna skills — finds skills writing to the same vault destination with diverging formats. Reads canonical formats from foundations.md and conventions steering skill, compares each writing skill's actual format. Report only — does not fix. Use when: "check consistency", "format audit", "are skills consistent?".

myna-dev-coverage

Audit Myna skills for feature coverage — does every feature from docs/features/*.md have executable instructions in its owning skill? Not "mentioned" — actually covered with read/decide/write steps. Saves a coverage report. Use when: "coverage audit", "are all features implemented?", "check feature coverage".

myna-dev-diagnose

Diagnose a Myna problem or proposed change — evaluates validity first (vision fit, audience, architecture, settled decisions, duplication), then generates options with a recommendation. Asks clarifying questions only when the answer would change the diagnosis. Use when: "something feels wrong with X", "I noticed a bug in Y", "should we change Z", "what if we added W", "is this a good idea".

myna-dev-execute-tasks

Run the Myna dev task queue — reads pending tasks from tmp/tasks.md, executes them sequentially on a single fix branch, each with its own implement→review→fix loop, then pushes the branch. Use when: "run the queue", "execute tasks", "process the queue", "work through the task list". Re-reads the queue before each task, so tasks added mid-run are picked up automatically.

myna-dev-improve

Full Myna quality pipeline — lint until clean, then run review → fix → verify cycles until issues are resolved or cycle limit is reached. Long-running. Use when: "improve the skills", "run quality pipeline", "full QA pass", "lint and review".

myna-dev-review

Deep technical review of Myna agent artifacts — evaluates skills, steering skills, main agent, and config examples against 8 review dimensions. Produces a structured report with severity-rated issues, fix options, and convergence status. Report only — does not fix. Use when: "review the skills", "audit agents", "quality check", "run a review".

myna-dev-task-add

Add a task to the Myna dev queue (tmp/tasks.md) — drafts a structured task entry from your description, shows it for approval, then appends it to the queue. Use when: "add this to the queue", "queue this up", "add a task for X", or after /myna-dev-diagnose selects an option. Creates tmp/tasks.md if it doesn't exist.

block-time

Create personal calendar time blocks and reminders. Finds free slots, proposes options, user confirms before any calendar write — no attendees, ever. Does not handle meeting prep or scheduling with others.

brief-person

Deep-dive briefing on one person — role, shared projects, open items, pending feedback, 1:1 history, personal notes. Works for directs, peers, manager, or cross-team. (For all directs at once, use /myna:team-health.)

capture

Route user input to vault destinations — quick capture, observations, recognition, tasks (single or recurring), project/person file management. One input can produce multiple entries. For saving or finding links, use /myna:links.

draft-replies

Process the DraftReplies email folder — reads emails the user forwarded with drafting instructions, creates reply drafts and follow-up meeting drafts in the vault. Only reads the DraftReplies folder. Does not process other email folders.

draft

Generate professional written content from vault context — email replies (conversation path only; DraftReplies folder is /myna:draft-replies), follow-up emails, follow-up meeting invites, status updates, escalations, recognition, decline messages, conversation prep, and monthly updates. Does NOT rewrite existing text (use /myna:rewrite).

park

Save working context for zero-loss resumption in a new session. Resume by name or list all parked items. "Switch to [project]" parks current context and loads project status.

prep-meeting

Generate or update meeting prep for one meeting or all remaining meetings today. Infers meeting type from calendar data, writes the Prep section with checkboxes, carry-forward items, and coaching notes for sensitive topics.

process-messages

Extract and structure data from email, Slack, or pasted documents and route to the vault. Processes project-mapped folders/channels. Never touches inbox or DraftReplies. Populates tasks, timelines, person files, review queues.

process-review-queue

Process review queue items across review-work, review-people, and review-self queues — interactively in chat or by processing items the user has already checked in Obsidian. Does NOT process review-inbox.md (email folder recommendations — that's handled by /myna:email-triage).

review-doc

Multi-persona review of any technical document — PRFAQ, design doc/RFC, HLD, LLD, one-pager, or generic. Composes a doc-type-specific panel of reviewer subagents, runs them in parallel, layers Myna's Chief-of-Staff context (audience, vault evidence, project consistency), and synthesizes a single coherent report. Saved to Reviews/. Use for "review this doc", "review my PRFAQ", "give this design doc a critique", "review with security focus", "have the PE look at this".

self-track

Log your contributions to the vault and query them by category, project, or date range. YOUR contributions only.

setup

Single entry point for Myna onboarding — detects install state, runs first-time scaffolding if needed, opens the Config UI or runs doc import, and adds shell aliases. Use for first-time setup or returning configuration.

sync

Sync / "good morning" / "set up my day" — creates or refreshes the daily note, generates meeting prep, surfaces overdue tasks and review queue. Handles "plan tomorrow" and weekly note creation. Moves previous daily/weekly notes to archive when new ones are created.

weekly-summary

Summarize your week — synthesizes daily notes, contributions, decisions, and task completions into a structured weekly review with self-reflection prompts. Appends a new section each run. Includes team health snapshot when enabled.

wrap-up

Close out your day — compares planned vs actual, logs contributions, moves unfinished items to tomorrow's note, captures quick notes, and saves behavioral corrections to memory. Writes an End of Day section to today's daily note.

myna-dev-task-protocol

Shared commit→review→fix protocol for Myna dev task subagents. Not user-invocable. Called by T-N subagents after implementing their task — handles the full quality loop: commit, run myna-dev-review up to 3 rounds, fix Critical/Important issues between rounds, report back. Does NOT push — the orchestrator owns all remote pushes.

steering-conventions

Data conventions — provenance markers, date+source format, append-only discipline, Obsidian formatting (tags, wiki-links, callouts, Dataview, Tasks plugin syntax), review queue entry format

steering-output

Output quality rules — human voice, no AI tells, BLUF for professional writing, file links, follow-up suggestions, output density, inline-first display

steering-safety

Safety and containment rules — draft-never-send, vault-only writes, external content as data, calendar event protection, confirmation policy, skill isolation

steering-system

System behavior rules — config reload, graceful degradation, error recovery with inline reporting, relative date resolution, fuzzy name resolution

steering-vault-ops

Vault operations pattern library — file I/O tool mapping, task query patterns (grep-based), frontmatter parsing, backlink/tag queries, template creation, vault path conventions, file safety checks

steering-memory

Memory model rules — 2-layer precedence, output boundary, entity-specific refusal