Issues

MemoryTurnStart sends the whole hot layer as one hook output; once memory fills, Claude Code cuts it to a 2KB preview and most memory never reaches the model

#2103

· Nkburdick opened

on Sep 11, 2026

Pulse serves stale code indefinitely after a source change, and cannot exit cleanly to reload

#2101

· bowlax opened

on Sep 11, 2026

Five .env parsers disagree: two return an empty env on CRLF files, three mis-key an export prefix

#2100

· bowlax opened

on Sep 11, 2026

The public-push check can miss the branch you're pushing

#2099

· hjbrandt opened

on Sep 11, 2026

Memory health permanently WARNs on memory-retrievals.jsonl: CortexHealth checks a stream MemoryRetriever never writes

#2093

· dth0m opened

on Sep 10, 2026

Capture.sh collapses eight distinct preflight exit codes to a constant 3 against a header (and an in-file comment) promising propagation

#2089

· Steffen025 opened

on Sep 10, 2026

Interceptor Reproduce.md step 3 "Check Console Errors" emits consoleCheck: done from a probe that never reads the console

#2088

· Steffen025 opened

on Sep 10, 2026

Safety hook auto-allows mcp__* calls the operator listed in permissions.ask, resolving the prompt that entry exists to raise

#2087

· Steffen025 opened

on Sep 10, 2026

ConfigSystem.md documents settings.json as generated and never-edit while the shipped install deliberately never establishes the split; hand-creating settings.system.json silently flips which file is authoritative

#2086

· Steffen025 opened

on Sep 10, 2026

parseGoals buckets by section heading, so on the shipped TELOS template no goal can ever leave "Active Goals"

#2085

· MatiasBarboza opened

on Sep 10, 2026

`memory-retrievals.jsonl` is graded by health but never written by anything

#2081

· aveneri opened

on Sep 9, 2026

memory-retrievals.jsonl is read by three health checks and written by nothing, so retrieval freshness fails permanently on a healthy install

#2080

· michaelaye opened

on Sep 9, 2026

Navigation Menu

Uh oh!

All issues

MemoryTurnStart sends the whole hot layer as one hook output; once memory fills, Claude Code cuts it to a 2KB preview and most memory never reaches the model

Pulse serves stale code indefinitely after a source change, and cannot exit cleanly to reload

Five .env parsers disagree: two return an empty env on CRLF files, three mis-key an export prefix

The public-push check can miss the branch you're pushing

Memory health permanently WARNs on memory-retrievals.jsonl: CortexHealth checks a stream MemoryRetriever never writes

Capture.sh collapses eight distinct preflight exit codes to a constant 3 against a header (and an in-file comment) promising propagation

Interceptor Reproduce.md step 3 "Check Console Errors" emits consoleCheck: done from a probe that never reads the console

Safety hook auto-allows mcp__* calls the operator listed in permissions.ask, resolving the prompt that entry exists to raise

ConfigSystem.md documents settings.json as generated and never-edit while the shipped install deliberately never establishes the split; hand-creating settings.system.json silently flips which file is authoritative

parseGoals buckets by section heading, so on the shipped TELOS template no goal can ever leave "Active Goals"

`memory-retrievals.jsonl` is graded by health but never written by anything

memory-retrievals.jsonl is read by three health checks and written by nothing, so retrieval freshness fails permanently on a healthy install

Uh oh!

Search results