Context
The triage scan workflow (#102) was motivated by a measured inefficiency: triaging 8 open issues required 11+ API calls, ~40 KB of tokens, and ~4 minutes of agent time. After implementing triage envelope fields and lifecycle labels, the same operation takes 2 API calls, ~400 B of parsed data, and <1 second.
These before/after numbers are concrete, reproducible, and demonstrate shiplog's value proposition for agent-driven workflows. The README should surface this — it's the kind of metric that helps users decide whether shiplog is worth adopting.
Design Summary
Approach
Single edit to README.md — add a benchmark section. Place it after the feature overview and before installation/setup, so it serves as motivation before the user commits to installing.
Benchmark Data (from #102 session)
| Metric |
Without triage fields |
With triage fields |
| API calls to triage 8 issues |
11+ (individual body fetches) |
2 (1 bulk list + 1 PR cross-ref) |
| Token input to agent |
~40 KB |
~400 B (envelope data only) |
| Wall time |
~4 minutes |
<1 second |
| Readiness visible at list level |
No |
Yes (lifecycle labels) |
| Task completion visible without body read |
No |
Yes (envelope triage fields) |
| Label-only filter ("show me ready issues") |
Not possible |
gh issue list --label shiplog/ready |
Tasks
Authored-by: claude/opus-4.6 (claude-code)
Captain's log entry created by shiplog
Context
The triage scan workflow (#102) was motivated by a measured inefficiency: triaging 8 open issues required 11+ API calls, ~40 KB of tokens, and ~4 minutes of agent time. After implementing triage envelope fields and lifecycle labels, the same operation takes 2 API calls, ~400 B of parsed data, and <1 second.
These before/after numbers are concrete, reproducible, and demonstrate shiplog's value proposition for agent-driven workflows. The README should surface this — it's the kind of metric that helps users decide whether shiplog is worth adopting.
Design Summary
Approach
Single edit to
README.md— add a benchmark section. Place it after the feature overview and before installation/setup, so it serves as motivation before the user commits to installing.Benchmark Data (from #102 session)
gh issue list --label shiplog/readyTasks
[tier-3]README.mdwith the triage scan benchmark table and a brief paragraph explaining what the numbers mean. Check the current README structure to determine the best placement.README.md(modify)Authored-by: claude/opus-4.6 (claude-code)
Captain's log entry created by shiplog