Brag doc & perf review

Standups capture today. Brag docs capture a quarter. DevRecall builds both from the same underlying activity store.

Brag doc

devrecall brag Q1-2026

Output is a Markdown brag doc structured for self-reviews and perf-review evidence:

## Q1 2026 Highlights

### Key Deliverables
- **Payment system rewrite** — led end-to-end. 47 commits, 8 PRs,
  reduced checkout latency by 40% (PROJ-89)
- **Auth service migration** — zero-downtime migration. 23 commits
  across 3 repos (PROJ-123, PROJ-124, PROJ-125)

### Collaboration & Mentorship
- Reviewed 34 PRs from 8 teammates
- Mentored Anna on payment integration — 12 PR reviews
- Led 3 design reviews for notification system

### Metrics
- 156 commits across 4 repos
- 23 PRs merged, 34 PRs reviewed
- 89 Slack threads in technical channels
- 47 meetings (32h total)

Flags:

devrecall brag Q1-2026 --format markdown   # default
devrecall brag Q1-2026 --format pdf        # via pandoc if installed
devrecall brag --since 2026-01-01 --until 2026-03-31
devrecall brag --output brag-q1.md

Perf review

devrecall perf-review Q1-2026

Same data, but framed for promotion / perf-review narratives: strong areas, areas to highlight, and concrete evidence.

✅ Strong areas
- Technical depth — auth, payments, infra
- Cross-team collaboration — backend + mobile + payments
- Consistent delivery — every sprint shipped

⚠ Areas to highlight
- More documentation / RFCs
- Ownership of larger initiatives

📝 Evidence
- Payment rewrite → measurable latency impact
- Auth migration → technical leadership
- 34 PR reviews → mentorship signal

How it builds

Brag docs use periodic summaries (daily → weekly → monthly → quarterly) that DevRecall generates automatically as you accumulate activity:

4,827 activities → 90 daily summaries → 13 weekly → 3 monthly → 1 quarterly

The quarterly summary is the brag doc’s spine; the LLM expands it with drill-down into specific deliverables and metrics.

Without an LLM

devrecall brag --raw produces an unsynthesized list of activities, metrics, and tickets touched in the period — useful as raw input to write the doc yourself.

Using a better model just for brag docs

Brag docs benefit from a strong LLM. Configure model routing per task:

{
  "llm": {
    "provider": "ollama",
    "model": "gemma4",
    "models": {
      "standup": "gemma4",
      "chat": "gemma4",
      "brag": "claude-sonnet-4-6"
    }
  }
}

Daily standups stay free on local Ollama; the quarterly brag doc flips to BYOK Claude for output quality.