Public: anyone can read this; nobody can change it.
workbenchthis machine
RBAC-5: seal a per-daughter Forgejo-webhook HMAC secret (plat/mitosis#32) #17
Closed without the crew
commented
Observability v1 SHIPPED this session: agents 0.7.x posts a single edit-in-place 🔄 working… heartbeat comment (elapsed/steps/last-action) on the PR/issue during a run, plus the idle-based uncapped timeout (long runs live as long as they emit; killed only when genuinely idle). This issue tracks the FULL version (streaming the live transcript into the Forgejo Actions UI). Design/options in the issue body; recommended path = run the agent AS a Forgejo Actions job so its stream-json stdout streams to the native job-log UI.
Sign in to join this conversation.
new above
offline
RBAC-5 (plat/mitosis#32): seal the webhook HMAC secret per daughter
fork_regen_secretsnow mints a freshWEBHOOK_SECRET(openssl rand -hex 32) and seals it into the daughter'splat-agents-secrets, sealed to the sovereign fork key. The agents Deployment already loads it viaenvFrom(plat-agents-secrets), so the dispatcher readsWEBHOOK_SECRETand both signs the Forgejo system-hook and verifiesX-Hub-Signature-256on/hookbefore dispatch (agents PR).Edits
fork_regen_secretsonly (germinate seam; bin/up untouched — DOM-2 owns it).Validation. Ran
fork_rekey_sops+fork_regen_secretsagainst a fresh gitops clone with a throwaway fork key: the decryptedplat-agents-secretscarriesWEBHOOK_SECRET(sops-decryptable by the fork key), and two independent runs produce distinct secrets (per-daughter randomness confirmed).Pairs with agents PR (verify + seal-into-Forgejo). Do not merge — orchestrator gates.
🤖 Generated with Claude Code