Strategy Skills

Braindump

Turn messy thoughts into organized vault entries.

Workflow

1. Receive raw input — Accept any unstructured text: stream-of-consciousness, voice transcript, bullet points, mixed topics, rambling notes. Do not ask the user to organize it first.

2. Extract discrete topics — Parse the input and identify distinct topics, ideas, or threads. Each becomes a separate note or gets appended to an existing note.

3. Classify each topic — Determine the best vault location:

• Project idea → projects/ideas/
• Task / action item → extract as task with - [ ] checkbox, link to relevant project
• Decision or opinion → decisions/ with decision-log format
• Learning / insight → references/ or relevant project's docs/
• Personal reflection → journal/
• Technical note → references/technical/
• Meeting note → meetings/

4. Generate backlinks — For each note, identify:

• Existing vault notes it should link to (use [[wikilink]] syntax)
• Tags that connect it to broader themes (#project-name, #idea, #decision)
• Any people mentioned → [[People/Name]]

5. Write frontmatter — Each note gets:

   ---
   type: braindump
   source: [user-input | voice-transcript | meeting]
   date: [YYYY-MM-DD]
   topics: [list of extracted topics]
   tags: [relevant tags]
   ---

6. Present summary — Show the user what was created:

• List of notes created/updated
• Key backlinks established
• Any action items extracted

Output Format

## Braindump Processed

**Input**: [brief description of what was provided]
**Date**: [YYYY-MM-DD]

### Notes Created
1. `projects/ideas/[topic].md` — [one-line summary]
2. `references/technical/[topic].md` — [one-line summary]

### Action Items Extracted
- [ ] [action] → linked to [[Project]]
- [ ] [action] → linked to [[Project]]

### Backlinks Established
- [[Existing Note]] ← new connection from [topic]

Vault Paths

Default vault: ~/broomva-vault/

If the vault is not accessible, output the notes as markdown blocks the user can manually save.

Tips

• Preserve the user's voice — don't over-formalize raw thoughts
• When in doubt about classification, use inbox/ and tag for later sorting
• Extract dates mentioned in the text and convert to absolute dates
• If the input mentions people, create or link to People notes

strategy-skills

Strategy & Decision Intelligence — 9 agent skills for the Broomva Stack (Layer 7).

Install

npx skills add broomva/strategy-skills

Skills

• pre-mortem — Structured risk analysis assuming project failure (4-category scoring, mitigation plan)
• premortem — Klein/Kahneman premortem with parallel sub-agent deep-dives + HTML report (adapted from @itsolelehmann)
• braindump — Capture raw thoughts → Obsidian vault
• morning-briefing — Daily brief from vault priorities
• drift-check — Priority alignment vs actual effort
• strategy-critique — Red-team strategy documents
• stakeholder-update — Multi-audience communication generator
• decision-log — Structured decision capture → vault
• weekly-review — Weekly change summary and attention flags

On pre-mortem vs premortem

These are two separate skills with different shapes:

• pre-mortem (the original) is structured-spreadsheet shaped: 4 categories (Technical / Execution / Market / Organizational), Likelihood × Impact scoring, ranked risk table with owners and deadlines. Best for org-level strategy reviews where you need a defendable risk register.
• premortem (this addition) is narrative shaped: parallel sub-agent deep-dives produce failure stories one per reason, with underlying assumptions and early warning signs surfaced explicitly. Renders an HTML report. Best for individual decisions where the value is in surfacing hidden assumptions, not in producing a register.

Use whichever fits the decision. They compose: run premortem to surface the assumptions, then pre-mortem to score and assign mitigations.

Attribution for premortem

The premortem skill is adapted from @itsolelehmann's AI-skill formulation, posted to X on 2026-05-02 as part of his Skillstack work. The underlying method is Gary Klein's premortem technique (Harvard Business Review, 2007), popularized by Daniel Kahneman in Thinking, Fast and Slow. See .skills/premortem/SKILL.md for the full attribution chain.

Part of bstack

This is Layer 7 of The Broomva Stack.

License

MIT — Carlos Escobar / BroomVA