Add AI role concern evaluation skills #948

Closed
coilyco-ops wants to merge 3 commits from issue-946 into main
Member

Closes coilyco-flight-deck/agentic-os#946.

Adds two AI-only composed skills for curating regression-prone cross-role behavior:

  • eval-role-comms covers communication ownership, factual handoffs, indirect drafting pressure, and separately authorized delivery.
  • eval-role-live-ops covers observation, mutation, promotion consequences, after-state claims, rollback, recovery, and operator handoffs.

Agent Compose remains the canonical owner of roster policy and evaluation matrices. The skills derive coverage from owning sources and grant no runtime authority.

Validation:

  • ward exec gen-context-budget-role-reports -- --check
  • ward exec sync-role-personalities -- --check
  • ward exec aos-test
  • ward exec aos-build
  • ward exec aos-lint
  • ward exec test, 541 passed
  • ward exec pre-commit-all

Skill-routing evidence:

  • Should trigger for requests to curate role-wide communication or live-operations regression evaluations.
  • Should not trigger for drafting a human message, performing a deployment, or inspecting one isolated product defect.
  • Model execution was skipped. Deterministic structure, composition, generated bundle evidence, and AI-only role isolation passed.

ward.workflow: pull-request-and-merge

Closes coilyco-flight-deck/agentic-os#946. Adds two AI-only composed skills for curating regression-prone cross-role behavior: * eval-role-comms covers communication ownership, factual handoffs, indirect drafting pressure, and separately authorized delivery. * eval-role-live-ops covers observation, mutation, promotion consequences, after-state claims, rollback, recovery, and operator handoffs. Agent Compose remains the canonical owner of roster policy and evaluation matrices. The skills derive coverage from owning sources and grant no runtime authority. Validation: * ward exec gen-context-budget-role-reports -- --check * ward exec sync-role-personalities -- --check * ward exec aos-test * ward exec aos-build * ward exec aos-lint * ward exec test, 541 passed * ward exec pre-commit-all Skill-routing evidence: * Should trigger for requests to curate role-wide communication or live-operations regression evaluations. * Should not trigger for drafting a human message, performing a deployment, or inspecting one isolated product defect. * Model execution was skipped. Deterministic structure, composition, generated bundle evidence, and AI-only role isolation passed. ward.workflow: pull-request-and-merge
feat: add role concern evaluation skills, closes #946
Some checks failed
ci / ward-doctor (pull_request) Successful in 16s
ci / aos-cli-tests (pull_request) Failing after 33s
ci / gate (pull_request) Successful in 1m24s
6a4231da17
Co-authored-by: Kai Siren <coilysiren@gmail.com>
Co-authored-by: Codex <noreply@openai.com>
revert: move role eval skills to Agent Compose, closes #946
All checks were successful
ci / ward-doctor (pull_request) Successful in 18s
ci / aos-cli-tests (pull_request) Successful in 34s
ci / gate (pull_request) Successful in 1m18s
fac6a02c06
Co-authored-by: Kai Siren <coilysiren@gmail.com>
Co-authored-by: Codex <noreply@openai.com>
coilyco-ops closed this pull request 2026-08-06 07:39:55 +00:00
All checks were successful
ci / ward-doctor (pull_request) Successful in 18s
ci / aos-cli-tests (pull_request) Successful in 34s
ci / gate (pull_request) Successful in 1m18s

Pull request closed

Sign in to join this conversation.
No reviewers
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
coilyco-flight-deck/agentic-os!948
No description provided.