Cs Capture

/cs:capture — Brain-Dump Organizer

Command: /cs:capture <dump-text-or-path>

The cs-capture persona organizes a dump into 4 actionable sections with zero information loss.

When to Run

• You have an unstructured block of mixed thoughts to organize
• You need workspace connections surfaced (Glob+Grep verified)
• You want concrete next-action offers, not generic "consider X" suggestions

The skill ALSO triggers automatically without /cs:capture when you:

• Use trigger phrases like "brain dump", "let me dump some ideas", "here's everything on my mind", "I need to organize my thoughts"
• Paste a long unstructured block of mixed ideas (implicit trigger)

/cs:capture is the explicit form — useful when your dump doesn't include the trigger phrasing but you still want the organize behavior.

What You Get

For a typical 8+ item dump (4-section format):

## Projects & Ideas
{clustered themes with embedded questions/decisions}

## Tasks
{flat scannable action list}

## Connections
{real workspace links — Glob+Grep verified, never fabricated}

## How I Can Help
{concrete offers — what + where}

**Which of these should I tackle?**

For a small dump (≤5 unrelated items, compressed format):

## What I heard
- ...

## How I can help
- ...

Which should I tackle?

Discipline

• Capture everything — zero loss; trivial items go in; user prunes later
• Preserve voice — no corporate-ifying; "build something crazy with AI" stays as that
• Match output to input — small dumps get compressed format, not forced 4 sections
• No fabrication — Section 3 only surfaces real workspace matches
• No action without pick — only auto-action is the organization itself
• Max 1 clarifier per dump — only when one item is genuinely ambiguous between task and project

Workflow

# 1. (Optional) Pre-classify the dump as a heuristic seed
python ../skills/capture/scripts/dump_classifier.py path/to/dump.txt

# 2. (Optional) Recommend output format
python ../skills/capture/scripts/complexity_estimator.py path/to/dump.txt

# 3. Inventory the workspace for Section 3 candidates
python ../skills/capture/scripts/workspace_inventory.py \
  --root . --keywords "<extracted-keywords-from-dump>"

# 4. Persona organizes + delivers four (or compressed) sections.
# 5. Persona ends with "Which of these should I tackle?" and waits.

Stop Conditions

• Four sections (or compressed) delivered → done with the auto-action portion
• User picks a Section-4 offer → execute that offer
• User says "go" without picking → honor it but flag any items you weren't sure about

Anti-Patterns Rejected

• Fabricating workspace connections that weren't actually verified
• Dropping items deemed "trivial"
• Corporate-ifying the user's casual language
• Forcing 4-section structure when input is small
• Acting on Section-4 offers immediately without approval
• Splitting decisions/questions into separate top-level categories instead of embedding them
• Vague Section-4 offers ("you might want to consider…")

Related

• Agent: cs-capture
• Skill: capture
• Source spec: megaprompts/05-capture-megaprompt.md
• Adjacent commands: /cs:grill-me (slow deliberate plan grill), /cs:grill-with-docs (docs-anchored grill), /cs:handoff (session continuation)

Version: 1.0.0 Source: Path-B direct conversion of megaprompts/05-capture-megaprompt.md