Capture

Capture

A reliable front door for every commitment, idea, and open loop — and a processing pass that empties it. Capture is cheap and fails silently: you never notice the thought you didn't record, only the dropped commitment weeks later. This skill covers the capture conventions, the triage decision rules that convert fragments into executable next actions, and the weekly pass that keeps the inbox at zero.

When to use this skill

• Commitments keep slipping and you are discovering them from other people's follow-ups
• Your notes app, task inbox, and paper notebook have all become backlogs nobody drains
• The same items get reviewed week after week and never get done
• You are rebuilding a productivity system that collapsed and want the load-bearing part first
• Captured items are phrased as topics (pricing, taxes) rather than actions
• You want to measure whether the capture habit is actually working, not assume it

Inputs the skill expects

• A capture log — JSON with text and ideally captured_at, or a markdown bullet list
• The set of surfaces you currently capture into (phone, notebook, email, chat)
• Which single place you intend to use as the processing inbox
• Your current weekly processing slot, if one exists
• A reference date for staleness analysis (passed explicitly — the tools never read the clock)

Clarify First

Before generating, confirm these inputs. If any is unknown or vague, ASK — do not assume:

• Which inbox is the processing inbox — the whole design collapses if items drain to more than two places
• Whether the log carries capture timestamps — without them, staleness and habit-health analysis are unavailable and the audit degrades to phrasing only
• Whether a weekly processing slot already exists — determines whether this is a setup job or a repair job
• Tolerance for dropping items — sets how aggressive the drop recommendations are; some users need permission, others need a brake

Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.

Workflows

Workflow 1 — Triage the inbox into buckets with next actions

Run this at the start of every weekly pass. It does the mechanical sorting so your judgement is spent on the genuinely ambiguous items.

1. Drain every capture surface into one file (JSON or markdown bullets).
2. Run the triage tool with today's date so ageing is computed.
3. Do the flagged 2-minute items immediately, before filing anything else.
4. Rewrite every item flagged no-verb or unactionable-phrasing using the grammar <concrete verb> <specific object> <qualifier>.
5. Move each remaining item to its destination — actions, projects, calendar, waiting-for, reference, someday, or the bin.

python3 personal-productivity/capture/scripts/capture_triage.py \
  --input personal-productivity/capture/assets/sample_capture_log.json \
  --today 2026-07-21

Filter to one bucket when working a single destination at a time:

python3 personal-productivity/capture/scripts/capture_triage.py \
  --input personal-productivity/capture/assets/sample_capture_log.json \
  --today 2026-07-21 --bucket project --format json

Workflow 2 — Audit whether the capture habit is real

Monthly. Answers a different question from triage: not "what is in the inbox" but "is this system actually working."

1. Run the audit against the full log with a reference date.
2. Read the status line — healthy, at-risk, or failing — and the verdict, which names the dominant failure (processing vs phrasing).
3. Check the source breakdown for the leakiest capture surface.
4. Fix exactly one thing before the next month: either restore the processing pass, or fix the phrasing habit. Not both — they need different attention.

python3 personal-productivity/capture/scripts/capture_audit.py \
  --input personal-productivity/capture/assets/sample_capture_log.json \
  --today 2026-07-21

Tighten the staleness threshold if you process more than weekly:

python3 personal-productivity/capture/scripts/capture_audit.py \
  --input personal-productivity/capture/assets/sample_capture_log.json \
  --today 2026-07-21 --stale-days 7 --format json

Workflow 3 — Run the weekly inbox-zero pass

The load-bearing habit. 45 minutes, same slot weekly, defended like a meeting with someone you respect.

1. Open assets/weekly-processing-pass.md and work down it.
2. Gather from every surface (5 min), then triage (5 min).
3. Process every item one at a time without skipping (20 min) — skipping is how items become permanently stale.
4. Review projects for missing next actions (8 min), then waiting-for items (5 min).
5. Record the outcome counts on the checklist. Watch the drop rate specifically: below 15% means you are filtering too early at capture time.

mkdir -p build
python3 personal-productivity/capture/scripts/capture_triage.py \
  --input personal-productivity/capture/assets/sample_capture_log.json \
  --today 2026-07-21 --format json > build/triaged.json

Decision frameworks

The processing sequence

Apply in this order for every item. The order prevents the most common error — doing a 90-second task that should have been deleted.

Do / defer / delegate / drop

Drop is the most under-used outcome and the highest-leverage one. A healthy pass drops 15-25% of items. Below that you are filtering at capture time, which uses exactly the tired, distracted judgement that capture exists to protect you from.

Habit-health thresholds