| name | ll-execution-loop-job-runner |
| description | Governed execution-runner skill for starting or resuming the ready-job loop through the canonical workspace runtime without turning repo CLI carrier commands into skill authority. |
LL Execution Loop Job Runner
This is the canonical governed skill bundle for the Execution Loop Job Runner. It is the user-facing skill authority for Claude/Codex. It does not create a second runner. It routes into the canonical workspace carrier at ll loop run-execution and ll loop resume-execution.
Canonical Authority
- Governing ADRs:
E:\ai\LEE-Lite-skill-first\ssot\adr\ADR-018-Execution Loop Job Runner 作为自动推进运行时.MD, E:\ai\LEE-Lite-skill-first\ssot\adr\ADR-020-标准 Skill 实现、Adapter 安装与 CLI 边界基线.MD
- FEAT:
E:\ai\LEE-Lite-skill-first\ssot\feat\FEAT-SRC-003-002__runner-用户入口流.md
- Canonical runtime command:
python skills/l3/ll-execution-loop-job-runner/scripts/runner_operator_entry.py invoke --request <request.json> --response-out <response.json>
- Real execution carrier:
python -m cli.ll loop run-execution --request <request.json> --response-out <response.json>
- Real resume carrier:
python -m cli.ll loop resume-execution --request <request.json> --response-out <response.json>
Runtime Boundary Baseline
- Interpret this skill via
E:\ai\LEE-Lite-skill-first\ssot\adr\ADR-038-运行时核心抽象边界与对象分层基线.MD.
- This bundle is a governed operator-facing
Skill wrapper over the execution-loop carrier. It does not reclassify loop commands, run traces, or job records into skill authority.
Required Read Order
ll.contract.yaml
input/contract.yaml
output/contract.yaml
agents/executor.md
agents/supervisor.md
input/semantic-checklist.md
output/semantic-checklist.md
Execution Protocol
- Accept one
skill.execution-loop-job-runner request envelope.
- Require
payload.runner_scope_ref and explicit payload.entry_mode of start or resume.
- Route
start to ll loop run-execution and resume to ll loop resume-execution.
- Preserve the authoritative runner context, runner run ref, and entry receipt produced by the loop carrier.
- Return a skill-scoped response envelope that points back to the delegated loop command and emitted runner artifacts.
- Do not ask the operator to relay downstream skills manually.
Workflow Boundary
- Input: one governed runner entry request
- Output: one governed runner entry response with runner context, receipt, and post-run snapshot refs
- Downstream handoff: execution runner queue consumption via the loop carrier
- Out of scope: direct job claim/run/fail verbs as user-facing skill authority, inventing a second runner implementation, and mutating queue state outside the canonical loop carrier
Non-Negotiable Rules
- Do not replace the canonical skill with
ll skill execution-loop-job-runner.
- Do not bypass the loop carrier with handwritten runner receipts or response files.
- Do not turn the skill into a manual dispatcher for downstream workflow skills.