← charleslikesdata.com afk cockpit
11 repos · 1,235 runs · 1,017 issues · 907 merged · data through Jun 7 – Sep 3, 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 = 930 gate-passed
98%
landed rate (executor repos)
n = 11 exec repos
$4.13
API-equiv $ / merged PR
n = 907 merged
fully loaded: includes failed runs, quarantines, and scout overhead
32%
quarantine (safety) rate
n = 1,017 issues
18%
escalation rate
n = 1,235 runs
0%
rejected rate
n = 2,582 proposals
Pipeline

Stage funnel

proposed
2,582
promoted
1,066
41% of proposed
attempted
1,017
95% of promoted
passed gate
930
91% of attempted
published
913
98% of passed gate
merged
907
99% of published
By repository

Outcomes per repo

landed quarantined other
Orbit-Wars
131 · 89%
afk-agent-system
502 · 90%
afk-cockpit
106 · 94%
breakroom
37 · 84%
graphmark
15 · 93%
household-bms
52 · 92%
household-command-center
2 · 50%
housing-commute-analysis
15 · 93%
kaggriculture
22 · 91%
ragmark
9 · 89%
the-vault
0 · 0%
the-workshop
126 · 80%
Producers

Proposal quality

dropped
335
harvest
5
mis-sized
167
pending
1007
rejected
1
shipped
1066
superseded
1
Safety axis

Quarantine by category

324 total
other
86
no-op
46
review
30
scope
29
capability
22
integration-conflict
20
build-crash
17
deny-surface
13
setup-crash
10
test-conflict
10
timeout
8
circuit-breaker
7
auth
6
coupling
4
environmental
4
idle-timeout
4
unrecoverable-signature
3
flaky-test
2
security
2
rate-limit
1
Retries & recovery

Runs per issue

1.21 avg runs/issue
12345
Retries & recovery

Quarantine → retry rescue

retry rescues 63% of first-run quarantines · 107% of those actually retried
75%
first-try pass rate
14%
final-state quarantine (issues)
26%
quarantine rate (runs)
Cost · API-equivalent
fleet telemetry · per attempt · Jun 7 – Sep 3, 2026

Spend by model tier

$3,061.46 total across the window
sonnet$2087.91 · avg $1.70/attempt · n=1226 attempts
unknown tier$670.16 · avg $1.69/attempt · n=397 attempts
opus$303.39 · avg $2.45/attempt · n=124 attempts
⚠ 22% of attempt spend ($670.16 · 397 attempts) has no recorded model_tier — telemetry gap, not a model
Overhead · orchestrator + scout ops
$681.82
686 cost events · additive to the tier spend above · not a model tier
Claude Code transcripts, afk worktrees only · per API call · Aug 5 – Sep 3, 2026
$0.04
median / call
$0.11
p90 / call
7,182
calls logged
whole machine, all projects: $5,885.79 across 36,477 calls
Cost dial

Per-API-call cost distribution

4224
2006
607
202
109
31
3
<$.05.05–.10.10–.15.15–.20.20–.30.30–.50$.50+
Model mix

Opus share by day

Aug 5 range: 0%–63% Sep 3
Throughput

Slices per day

1,235 runs
06-0706-1606-2707-0607-2308-0208-1108-2009-02
Context cost

Tokens by type, per day

666.5M tokens · 96.5% cache reads — the context is overwhelmingly cached
cache read cache write input output