Skip to content

feat: Add agent teams support for parallel skill execution#4

Open
lmiccini wants to merge 3 commits into
openstack-k8s-operators:mainfrom
lmiccini:feature/agent-teams-support
Open

feat: Add agent teams support for parallel skill execution#4
lmiccini wants to merge 3 commits into
openstack-k8s-operators:mainfrom
lmiccini:feature/agent-teams-support

Conversation

@lmiccini
Copy link
Copy Markdown

Add experimental agent teams support enabling parallel execution across four skills: /code-review (3 parallel reviewers), /feature (3 parallel researchers), /task-executor (parallel task groups in worktrees), and /debug-operator (parallel hypothesis testing).

New team-role agents: researcher (read-only analysis), implementer (write-capable task execution), and reviewer (focused code review with cross-validation). All skills gracefully fall back to sequential behavior when CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS is not enabled.

Includes shared infrastructure (lib/team-helpers.sh), 44 new tests (tests/test-teams.sh), CI integration, and design documentation.

lmiccini and others added 2 commits April 28, 2026 17:39
Add experimental agent teams support enabling parallel execution across
four skills: /code-review (3 parallel reviewers), /feature (3 parallel
researchers), /task-executor (parallel task groups in worktrees), and
/debug-operator (parallel hypothesis testing).

New team-role agents: researcher (read-only analysis), implementer
(write-capable task execution), and reviewer (focused code review with
cross-validation). All skills gracefully fall back to sequential
behavior when CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS is not enabled.

Includes shared infrastructure (lib/team-helpers.sh), 44 new tests
(tests/test-teams.sh), CI integration, and design documentation.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a 4th researcher teammate to /feature parallel research that
challenges the other researchers' findings. The devil's advocate
spawns after the initial 3 researchers report, receives all findings,
and critiques them for unsupported assumptions, missed alternatives,
risks, and convention gaps. Valid concerns are incorporated as risks
in the implementation strategies.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Luca Miccini <lmiccini@redhat.com>
@lmiccini lmiccini force-pushed the feature/agent-teams-support branch from 2e6fcfb to 9caf106 Compare April 28, 2026 16:44
New agent-backed skill for writing tobiko tests, AnsibleTest playbooks,
and generating test-operator CRs (Tempest, Tobiko, AnsibleTest,
HorizonTest). The agent encodes deep domain knowledge including CRD
schemas, tobiko fixture patterns, and the iha-tests playbook convention.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: Luca Miccini <lmiccini@redhat.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant