Julian lechnerandClaude Sonnet 5 4374c69caa fix(claude-agents): map agent capabilities to real Claude Code tool names ...

Search, Diff, Shell, and Tests are not valid Claude Code tools, so every
generated subagent silently launched with a broken or empty tool set. Map
each role to the tools its own prompt actually needs, and use a __SHELL__
placeholder so build resolves it to the target package's own shell tool
instead of granting both Bash and PowerShell unconditionally.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

2026-09-18 09:31:27 +02:00

..

claude

fix(claude-agents): map agent capabilities to real Claude Code tool names

2026-09-18 09:31:27 +02:00

codex

feat(codex): constrain sandbox and agent depth

2026-09-17 22:59:13 +02:00

plugins.tsv

feat(qmd): add managed extension integration

2026-09-17 11:45:41 +02:00

prompt-budget-baseline.json

perf(prompt): reduce permanent instruction context

2026-09-17 14:37:18 +02:00

prompt-budget-baseline.tsv

perf(prompt): reduce permanent instruction context

2026-09-17 14:37:18 +02:00

rule-skills.tsv

refactor: share rule skill manifest

2026-09-17 13:16:24 +02:00