# Decision Log

> This skill should be used when the user asks to "log this decision," "record why we decided X," or needs a decision and its rationale added to a persisted, append-only decision log for later audit or onboarding reference.

- Skill: `daddia/decision-log` (Agent Skill)
- Install (CLI): `npx skillmds@latest add daddia/decision-log`
- Raw SKILL.md: https://api.skillmd.com/api/skills/daddia/decision-log/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: daddia (https://skillmd.com/u/daddia)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/daddia/decision-log

---

# Decision Log

## When to use

Append decisions to persisted, append-only audit log — choice made, rationale, owner, revisit trigger.

## What this skill does not do

- **Does not edit past entries** — supersede with new entry referencing original date.
- **Does not log topics discussed** — requires precise decision statement.

## Preconditions

| Input | If missing |
|---|---|
| Decision statement | Ask — distinguish decision from topic |
| Rationale | Ask — required for audit value |
| Practice profile (format, location) | Use sensible default; flag `[PROVISIONAL]` |

## Provisional mode

Without rationale: halt and ask before appending.

## Trust spine

- **Confidence bands** (`governance-tracking`):
  - **High:** Decision, rationale, owner, date, revisit trigger captured; append-only preserved.
  - **Medium:** Revisit trigger "none specified" when user confirms none.
  - **Low:** Topic not decision — reject entry.
- **Failure modes:**
  - **Accountability gap:** Owner required; no anonymous decisions.
  - **Incentive Gaming:** N/A — audit trail focus.
- **Escalation triggers:** User asks to edit history — supersede pattern only.

## Workflow

1. Read practice profile for format and location.
2. Get current log — append only.
3. Capture decision precisely, rationale, owner, date, revisit trigger.
4. Output full updated log with NEW ENTRY block.
5. Gaming-pattern check: no retroactive edits.

## Output format

```
[Full decision log, existing entries unchanged]
NEW ENTRY:
Date: [date]
Decision: [precise statement]
Rationale: [why vs alternatives]
Owner: [who]
Revisit trigger: [trigger or none specified]
```

## Worked example

**Input:** "We decided Identity will be a shared platform service."

**Expected entry:** Decision precise; rationale asks if missing; owner recorded.

## Quality checks before delivering

- [ ] Decision not topic
- [ ] Rationale present
- [ ] Append-only — no edits to prior entries
- [ ] Owner named

## Propose profile update

When a stable convention surfaces during this run (thresholds, naming, tone, output format, or recurring corrections), **propose a profile update**: show the exact diff against `~/.claude/plugins/config/claude-for-strategy/pmo/CLAUDE.md` (org-wide facts go to `org-profile.md`), ask for confirmation, and write only on yes. Only `/pmo:practice-setup` auto-applies a full profile write.

## Outputs

Follows plugin `CLAUDE.md` § Outputs. Next: reference in `steering-pack` or onboarding brief.