Update Playbook

Evaluate the current conversation context against existing Packmind artifacts (standards, commands, skills) to identify what needs creating, updating, or deprecating. Produce a structured change report, then apply approved changes.

Process

Phase 1: Context Summary

Before dispatching any subagents, summarize the current conversation context. Extract:

• What was built or changed
• What patterns emerged or were established
• What decisions were made and why
• What mistakes were corrected
• What problems were solved

This summary is reused as input for all subsequent subagent prompts.

Phase 2: Triage (parallel, shallow)

Find which existing artifacts are relevant to the session. Read each references/triage-*.md file and construct one prompt per subagent:

## Conversation Context

<the context summary from Phase 1>

## Triage Task

<full contents of the corresponding references/triage-*.md file>

Launch all three as Task(general-purpose) subagents in a single message:

Each subagent returns a short list — no deep analysis, just relevance filtering.

Phase 3: Deep Analysis (parallel, focused)

Using the triage results, read each references/analyse-*.md file and construct one prompt per subagent. Include only the artifacts flagged as relevant by triage:

## Conversation Context

<the context summary from Phase 1>

## Relevant Artifacts

<the filtered list from the corresponding triage subagent>

## Analysis Task

<full contents of the corresponding references/analyse-*.md file>

Launch all three as Task(general-purpose) subagents in a single message:

Skip a subagent entirely if its triage returned no relevant artifacts and no new artifact ideas.

Phase 4: Consolidated Report

After all analysis subagents complete, consolidate their reports into a single change report presented to the user. Number every change sequentially so the user can selectively approve (e.g., "apply 1, 3, 5" or "all but 4"):

## Playbook Change Report

### Skill Updates (keep current)
1. [skill] <name>: <what changed and why>

### New Skills
2. [skill] <name>: <reason>

### Standard Updates
3. [standard] <name>: <what changed and why>

### New Standards (lintable rules only)
4. [standard] <name>: <reason>

### Command Updates (strong evidence only)
5. [command] <name>: <what changed and why>

### Deprecations
6. [standard|skill] <name>: <reason>

### No Changes Needed
- [standards|commands|skills]: <brief explanation if an artifact type needs no changes>

Ordering reflects priority: skill accuracy first, then standards, then commands. New commands are never proposed (command creation is user-driven).

Present this report and ask the user which changes to apply. Accept any of:

• All: apply every numbered change
• Inclusion list: "1, 3, 5" or "only 2 and 6"
• Exclusion list: "all but 4" or "everything except 2, 7"

Map the user's selection back to the specific artifact changes before proceeding to Phase 5.

Phase 5: Apply Changes

For each approved change from the Phase 4 report, edit the local installed files directly:

• Standards: .packmind/standards/<slug>.md
• Commands: .packmind/commands/<slug>.md
• Skills: .claude/skills/<skill-name>/ (SKILL.md, references/, etc.)

Edit only one instance of each artifact — packmind-cli diff compares the local copy against the server version.

For new artifacts, delegate to the corresponding creation skill (packmind-create-standard, packmind-create-command, packmind-create-skill).

Step 1: Preview

Run packmind-cli diff and present the output to the user. Verify the diff matches the intended changes — no unrelated modifications should be included.

If unrelated changes appear in the diff (e.g., from a previous session), inform the user and proceed only with the intended changes.

Step 2: Submit

Run packmind-cli diff --submit -m "<concise summary of all changes>" to submit the changes as proposals for human review on Packmind.

Step 3: Propagate

Ask the user whether they have validated the submitted changes on Packmind and wish to propagate them locally. If yes, run:

packmind-cli install --recursive

Reference Files