Code orchestration communication
I am looking to improve open claw’s communication around coding tasks it is orchestrating, how does this look for a plan?
Here’s a short prompt for Krill:
---
Implement a core outbound send interceptor (“execution-claim proof gate”) in OpenClaw.
Requirements:
- Runs in the one choke point before any channel adapter send (covers normal replies + tool sends + edits).
- Detect execution-claim language: “started / building / running / still working / in progress” (tuneable regex).
- Require at least one proof token line in the same message:
- Add config:
- Phase 1 default:
- Phase 2:
- Idempotency: prevent duplicate sends on retries using
- Add unit tests for: allow, rewrite, block, idempotent retry.
Deliver: code + tests + brief docs + recommended default config.
Here’s a short prompt for Krill:
---
Implement a core outbound send interceptor (“execution-claim proof gate”) in OpenClaw.
Requirements:
- Runs in the one choke point before any channel adapter send (covers normal replies + tool sends + edits).
- Detect execution-claim language: “started / building / running / still working / in progress” (tuneable regex).
- Require at least one proof token line in the same message:
SESSION:CMD:ARTIFACT:- Add config:
executionClaimGate.mode = warn|rewrite|block- Phase 1 default:
warnrewriteIDLE — not started yet. Next: <ask/unknown>- Phase 2:
block- Idempotency: prevent duplicate sends on retries using
emitKey = sha256(channel+target+normalizedText+timeBucket)- Add unit tests for: allow, rewrite, block, idempotent retry.
Deliver: code + tests + brief docs + recommended default config.
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
134,827Members
Resources
Similar Threads
Was this page helpful?
