← charleslikesdata.com afk cockpit
6 repos · 804 runs · 642 issues · 558 merged · data through Jun 7 – Jul 18, 2026

Autonomous-PR outcomes across the fleet. Cost is API-equivalent (notional on a Pro subscription). Quarantines are a safety signal — afk refused to ship — not failures.

Fleet state · not yet wired up
Live gate/promote/merge and dependency-frontier state isn't ingested yet — this panel is a placeholder for a future feature, not a fabricated number.
99%
landed rate
n = 566 gate-passed
99%
landed rate (executor repos)
n = 5 exec repos
excludes my-brain (10 quar)
$3.82
API-equiv $ / merged PR
n = 558 merged
fully loaded: includes failed runs, quarantines, and scout overhead
38%
quarantine (safety) rate
n = 642 issues
20%
escalation rate
n = 804 runs
0%
rejected rate
n = 962 proposals
Pipeline

Stage funnel

proposed
962
promoted
651
68% of proposed
attempted
642
99% of promoted
passed gate
566
88% of attempted
published
559
99% of passed gate
merged
558
100% of published
By repository

Outcomes per repo

landed quarantined other
Orbit-Wars
51 · 98%
afk-agent-system
320 · 99%
afk-cockpit
106 · 99%
claude-workflow
107 · 95%
household-bms
48 · 100%
my-brain
10 · 0%
Producers

Proposal quality

dropped
156
harvest
5
mis-sized
33
pending
115
rejected
1
shipped
651
superseded
1
Safety axis

Quarantine by category

244 total
other
80
no-op
40
review
19
capability
17
integration-conflict
17
build-crash
16
setup-crash
10
circuit-breaker
7
deny-surface
6
scope
6
test-conflict
6
timeout
6
idle-timeout
4
coupling
3
unrecoverable-signature
3
flaky-test
2
security
2
Retries & recovery

Runs per issue

1.25 avg runs/issue
12345
Retries & recovery

Quarantine → retry rescue

retry rescues 59% of first-run quarantines · 96% of those actually retried
71%
first-try pass rate
17%
final-state quarantine (issues)
30%
quarantine rate (runs)
Cost · API-equivalent
fleet telemetry · per attempt · Jun 7 – Jul 18, 2026

Spend by model tier

$1,943.65 total across the window
sonnet$1099.77 · avg $1.48/attempt · n=745 attempts
unknown tier$674.64 · avg $1.69/attempt · n=399 attempts
opus$169.24 · avg $2.53/attempt · n=67 attempts
⚠ 35% of attempt spend ($674.64 · 399 attempts) has no recorded model_tier — telemetry gap, not a model
Overhead · orchestrator + scout ops
$185.40
154 cost events · additive to the tier spend above · not a model tier
Claude Code transcripts, afk worktrees only · per API call · Jun 18 – Jul 18, 2026
$0.03
median / call
$0.07
p90 / call
34,662
calls logged
whole machine, all projects: $6,443.27 across 60,315 calls
Cost dial

Per-API-call cost distribution

27466
5758
1114
214
80
25
5
<$.05.05–.10.10–.15.15–.20.20–.30.30–.50$.50+
Model mix

Opus share by day

Jun 18 range: 0%–72% Jul 18
Throughput

Slices per day

804 runs
06-0706-1106-1506-1906-2506-2907-0307-1007-18
Context cost

Tokens by type, per day

2.37B tokens · 95.6% cache reads — the context is overwhelmingly cached
cache read cache write input output