Runner / Super Runner
Runner / Super Runner executes scoped work inside approved boundaries. It coordinates handoff, evidence capture, and escalation. It does not approve, verify itself, or expand scope.
Role type
Execution and coordination skill
Primary shift
Ad hoc task execution to bounded evidence-aware execution
Core output
Changed artifact or handoff packet with receipts
Boundary
Not Gate, final reviewer, approval authority, or self-verifier
AI-assisted execution can become fast but unsafe if scope, handoff, evidence, and claim boundaries are not explicit. Runner exists to make execution faster while preserving accountability, traceability, and escalation discipline.
Before
- ad hoc task execution
- unclear handoff boundaries
- speed without consistent receipt structure
- worker routing handled manually case-by-case
After
- scoped execution discipline
- evidence-before-claim behavior
- reproducible handoff packet creation
- stop-state detection
- escalation routing
- same-page role-detail explanation pattern
- readable evidence and non-claim presentation
runner-scoped-execution
Executes only the approved slice and avoids scope expansion.
runner-handoff-and-evidence-capture
Creates artifacts, receipts, validation notes, and missing-evidence disclosure.
super-runner-coordination-skill
Coordinates multi-step execution inside boundaries without becoming authority.
DORA / delivery performance
Used as a learning foundation for speed with stability, recovery, and delivery discipline.
Google SRE
Used as a learning foundation for stop-state, reliability, incident/recovery thinking, and operational discipline.
OpenTelemetry-style thinking
Used as an adapted reference model for trace, log, evidence, correlation, and receipt mindset.
Diataxis
Used as a learning foundation for handoff-readable documentation: how-to, reference, explanation, and tutorial-style guidance.
OWASP / NIST-style risk awareness
Used as an adapted learning foundation for boundary, security, data exposure, and claim-risk awareness.
Source-of-truth lock - scope boundary - task packet - worker routing - execution - validation evidence - missing-evidence disclosure - escalation or gate handoff - receipt.
Inputs
- owner intent
- source of truth
- allowed scope
- forbidden additions
- worker/task packet
- expected artifact
- validation requirement
Outputs
- changed artifact or handoff packet
- receipt
- route ledger
- telemetry/validation note
- missing evidence list
- non-claims preserved
- next safe action
Gate / Owner
approval, claim, release, or risk decisions are required
Checker / QA
deterministic validation and evidence review are required
Big Crew
architecture, implementation split, maintainability, handover, scalability, or cross-functional requirements appear
Surface Runner / Surface Guild
rendered proof, public-surface readability, or UI evidence is required
Available Evidence
- Runner / Super Runner local dedicated detail route exists.
- Provider-backed review exists only for the review/spec layer if applicable.
- Local validation receipts exist for the bounded implementation layer.
Missing Evidence
- No final A+ grade unless separately evidenced.
- No rendered proof unless accepted visual proof exists.
- No live verification unless deployed and verified.
- No production readiness claim.
Claims Not Made
- Runner is not Gate.
- Runner is not final reviewer.
- Runner is not approval authority.
- Runner does not verify itself.
- No A+ claim by default.
- No live verification claim.
- No production readiness or framework certification claim.
| Level | Label | Evidence to promote |
|---|---|---|
| 1 | Scoped Runner | can execute one bounded task without scope expansion |
| 2 | Evidence Runner | can produce receipt, validation summary, and missing-evidence list |
| 3 | Super Runner Coordination Skill | can coordinate multi-step execution while preserving authority boundary |
| 4 | Execution System Designer | can turn Runner pattern into reusable role-detail template and handoff discipline |