Route Keras fine-tuning and transfer-learning work through Mission Control with explicit baseline, unfreeze, and evaluation steps.
Use when Mission Control should recommend a read-only codebase intake burst with Repo Mapper, Test Finder, Docs Reader, Risk Scanner, and Dependency Mapper.
Show file and path ownership and conflicts in Mission Control. Use when the user wants to know which paths are locked, which agent owns them, which tasks are waiting, or how to…
Define memory, state, checkpoints, resumability, and retention policy for Mission Control workflows and agents.
Route 3D visual-regression and render-diff work through Mission Control with explicit evidence targets and scene-aware validation.
Use Mission Control to plan bounded TensorFlow model optimization work such as quantization or pruning with explicit tradeoffs.
Use Mission Control with NVIDIA AI-Q, NVIDIA Dynamo, and GPU diagnostics for deeper research and GPU-backed coding workflows.
Use when Codex should relay user intent to Mission Control instead of acting as the Manager itself.
Extract business domains, flows, process steps, and domain vocabulary from a codebase or knowledge base through Mission Control.
Design or audit tool registries, tool schemas, permissions, and runner capabilities through Mission Control.
Use when Codex should generate or review AGENTS.md through Mission Control context and explicit user approval.
Route OpenUSD and Houdini/Solaris workflow planning through Mission Control with explicit scene-graph, interchange, and validation boundaries.
Use when Codex should attach an existing repo or folder to Mission Control in read-first import mode.
Use when Codex should route a workspace task through Mission Control instead of acting like the Manager AI.
Coordinate brand-guideline, internal-communications, announcement, and stakeholder-message work through Mission Control with tone, audience, and evidence constraints.
Generate an operational runbook through Mission Control. Use when the user wants RUNBOOK.md or a chat-native runbook covering startup, tests, build, debugging, local reset, logs,…
Route CUDA kernel generation or repair through Mission Control with GPU-aware validation and infrastructure checks.
Run a post-handoff or mid-project Mission Control change request flow. Use when the user wants additional work classified, impact-estimated, task-created, validated, and folded…
Compare before-and-after performance evidence through Mission Control without turning benchmark claims into theater.
Request a Mission Control snapshot or restore point before risky work. Use when the user wants rollback safety before refactors, broad edits, or uncertain changes, preferably…
Route Blender headless render and validation work through Mission Control with explicit external-runtime blockers and evidence loops.
Use Mission Control to plan or repair TFX-style production ML pipelines with explicit component and validation boundaries.
Run a direct Mission Control fix workflow for an existing repo. Use when the task is a bugfix or targeted change in a non-empty codebase and the Manager should classify the…
Summarize recent Mission Control events without raw logs. Use when the user wants the last 5 minutes, last 15 minutes, since last interaction, or since orchestration start…
Skip the Mission Control interview and proceed with assumptions. Use when speed matters more than full intake and the user accepts that the Manager will proceed with explicit…
Check whether Mission Control claims have backing evidence. Use when the user wants to know whether tests, builds, screenshots, artifacts, or handoff confidence are real, missing,…
Use when Codex should attach an existing repo or folder to Mission Control and keep the first pass read-only until the user authorizes more.
Ask Mission Control Manager for recovery guidance after a failure. Use when the user wants a failure summary, likely causes, recovery options, recommendation, and risks without…
Design, build, or audit MCP servers and tool/resource/prompt contracts through Mission Control. Use for MCP schemas, stdio startup, auth boundaries, plugin packaging, and smoke…
Design graph-style agent workflows with nodes, edges, gates, retries, checkpoints, and human approvals through Mission Control.
Request compiler and autotuning guidance for GPU code through Mission Control with bounded iteration and evidence requirements.
Use when Mission Control needs a user approval or manager-question answer relayed through Codex chat.
Ask questions against Mission Control's codebase understanding. Use for architecture questions, where-is-this-handled, why-does-this-exist, and codebase Q&A grounded in the…
Use when Mission Control should recommend a read-only handoff audit burst for run instructions, validation evidence, limitations, docs quality, and security caveats.
Use when Mission Control should recommend a bounded read-only review burst across correctness, security, testing, maintainability, and docs.
Use when the user wants a compact Mission Control status summary, blocker report, or agent activity check.
Use Mission Control to turn TensorBoard logs and training curves into explicit validation evidence.
Use or prefer Claude CLI runner mode through Mission Control. Use when the user explicitly wants Claude CLI if configured and Codex should verify availability and fallback rather…
Continue work after a Mission Control handoff without losing continuity. Use when the user wants another iteration, wants limitations preserved, or needs the next change request…
Prepare a project for release through Mission Control. Use when validation, docs, versioning, changelog, limitations, evidence, deployment readiness, and security concerns need…
Ask Mission Control whether the project is drifting beyond approved scope. Use when the task list is growing, the user asks for more work midstream, or you need a Manager…