Skip to content

feat(conversation): agents talk + use tools [v0.3.0 — 3/7]#733

Merged
kokevidaurre merged 2 commits intodevelopfrom
v030/conversation
Apr 13, 2026
Merged

feat(conversation): agents talk + use tools [v0.3.0 — 3/7]#733
kokevidaurre merged 2 commits intodevelopfrom
v030/conversation

Conversation

@kokevidaurre
Copy link
Copy Markdown
Contributor

Summary — PR 3 of 7 for v0.3.0 release

Conversation mode rewrite and cognition engine.

Changes (2 files)

  • conversation.ts: Rewritten so agents talk AND use tools (was text-only). Parallel same-role agents. Hard-stop on lead completion. Squad cwd resolution. Agent classification by name first.
  • cognition.ts: Local-first intelligence engine. Quality grading. Escalation pause. Signal synthesis. Memory signal push after daemon cycles.

Merge order

Depends on #731 and #732. Merge third:

  1. ✅ core-refactor (refactor(core): run engine decomposition + context helpers [v0.3.0 — 1/7] #731)
  2. ✅ run-engine (feat(run): workflow rewrite — smart skip, org cycle, wave execution [v0.3.0 — 2/7] #732)
  3. → conversation (this PR)
  4. new-commands
  5. init-ux
  6. security-guardrails
  7. tests-docs

Test plan

  • npm run build passes
  • Review conversation state machine
  • Review cognition engine signal flow

Reorganized from 219-commit develop branch for proper review.

Copy link
Copy Markdown

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request introduces a significant architectural shift from a sequential, turn-based conversation model to a wave-based parallel execution workflow (Plan → Execute → Review → Verify). Key changes include the implementation of resume support for quota-limited runs, improved context injection strategies that prioritize actionable feedback, and the externalization of agent protocols into markdown configuration files. Feedback is provided regarding code duplication in the telemetry tracking logic within the run command, where a shared helper function could improve maintainability.

Comment thread src/commands/run.ts
…gence

Conversation mode rewrite and cognition engine from v0.3.0 cycle:

- conversation.ts: Rewritten so agents talk AND use tools (was text-only).
  Parallel same-role agents within cycles. Hard-stop on lead completion.
  Squad cwd resolution for all agent turns. Transcript serialization fixes.
  Agent classification by name first, then role description.
- cognition.ts: Local-first intelligence engine. Quality grading.
  Escalation pause for daemon. Signal synthesis via Claude CLI.
  Push memory signals after daemon cycles.

Co-Authored-By: Claude <noreply@anthropic.com>
Cognition engine is not actively used (post-pivot, daemon is stopped).
Changes parked in future/cognition-t2 branch for Tier 2 reactivation.
This PR now only contains conversation.ts changes.

Co-Authored-By: Claude <noreply@anthropic.com>
@kokevidaurre kokevidaurre changed the title feat(conversation): agents talk + use tools, cognition engine [v0.3.0 — 3/7] feat(conversation): agents talk + use tools [v0.3.0 — 3/7] Apr 13, 2026
@kokevidaurre kokevidaurre merged commit 922934f into develop Apr 13, 2026
11 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant