← charleslikesdata.com afk cockpit
6 repos · 810 runs · 646 issues · 572 merged · data through Jun 7 – Jul 25, 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.
98%
landed rate
n = 581 gate-passed
98%
landed rate (executor repos)
n = 5 exec repos
$3.78
API-equiv $ / merged PR
n = 572 merged
fully loaded: includes failed runs, quarantines, and scout overhead
37%
quarantine (safety) rate
n = 646 issues
20%
escalation rate
n = 810 runs
0%
rejected rate
n = 1,096 proposals
Pipeline

Stage funnel

proposed
1,096
promoted
679
62% of proposed
attempted
646
95% of promoted
passed gate
581
90% of attempted
published
573
99% of passed gate
merged
572
100% of published
By repository

Outcomes per repo

landed quarantined other
Orbit-Wars
53 · 98%
afk-agent-system
332 · 99%
afk-cockpit
106 · 99%
household-bms
48 · 100%
my-brain
0 · 0%
the-workshop
107 · 94%
Producers

Proposal quality

dropped
193
harvest
5
mis-sized
37
pending
180
rejected
1
shipped
679
superseded
1
Safety axis

Quarantine by category

236 total
other
80
no-op
32
review
19
capability
17
integration-conflict
17
build-crash
16
setup-crash
10
circuit-breaker
7
scope
7
deny-surface
6
test-conflict
6
timeout
5
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 62% of first-run quarantines · 98% of those actually retried
72%
first-try pass rate
16%
final-state quarantine (issues)
29%
quarantine rate (runs)
Cost · API-equivalent
fleet telemetry · per attempt · Jun 7 – Jul 25, 2026

Spend by model tier

$1,950.48 total across the window
sonnet$1129.71 · avg $1.48/attempt · n=763 attempts
unknown tier$648.07 · avg $1.70/attempt · n=382 attempts
opus$172.70 · avg $2.50/attempt · n=69 attempts
⚠ 33% of attempt spend ($648.07 · 382 attempts) has no recorded model_tier — telemetry gap, not a model
Overhead · orchestrator + scout ops
$214.44
186 cost events · additive to the tier spend above · not a model tier
Claude Code transcripts, afk worktrees only · per API call · Jun 25 – Jul 25, 2026
$0.03
median / call
$0.07
p90 / call
32,669
calls logged
whole machine, all projects: $7,425.96 across 63,805 calls
Cost dial

Per-API-call cost distribution

25909
5576
895
179
82
23
5
<$.05.05–.10.10–.15.15–.20.20–.30.30–.50$.50+
Model mix

Opus share by day

Jun 25 range: 0%–62% Jul 25
Throughput

Slices per day

810 runs
06-0706-1206-1706-2406-2907-0407-1407-23
Context cost

Tokens by type, per day

2.29B tokens · 95.7% cache reads — the context is overwhelmingly cached
cache read cache write input output