Use Cases

This document maps practical AGENT-33 deployments to the current runtime surface in engine/src/agent33.

1. Guardrailed Code Review Pipeline

Goal:

• Run scoped, repeatable review flows with explicit risk handling.

Use these modules:

• api/routes/workflows.py
• api/routes/reviews.py
• review/service.py
• evaluation/service.py

Typical flow:

1. Register a review workflow (/v1/workflows/).
2. Execute workflow for a PR branch (/v1/workflows/{name}/execute).
3. Create review record (/v1/reviews/).
4. Assess risk triggers and route L1/L2 signoff.
5. Record evaluation run and regressions for release gating.

Best fit:

• Teams that require explicit two-layer review before merge.

2. Release Command Center

Goal:

• Move releases through frozen -> RC -> validate -> publish with checklist controls.

Use these modules:

• api/routes/releases.py
• release/service.py
• release/checklist.py
• release/sync.py
• release/rollback.py

Typical flow:

1. Create release (POST /v1/releases).
2. Freeze, cut RC, validate.
3. Update/evaluate checklist items.
4. Publish when checks pass.
5. Run sync dry-runs or real syncs.
6. Initiate rollback if needed.

Best fit:

• Teams with repeatable release compliance requirements.

3. Autonomous Execution Budgeting

Goal:

• Enforce hard runtime limits for file, command, and network activity.

Use these modules:

• api/routes/autonomy.py
• autonomy/service.py
• autonomy/enforcement.py
• autonomy/preflight.py

Typical flow:

1. Create and activate budget.
2. Run preflight checks.
3. Attach runtime enforcer.
4. Gate each file/command/network action through enforcement APIs.
5. Trigger and resolve escalations.

Best fit:

• High-control automation in regulated or sensitive environments.

4. Evaluation and Regression Gates

Goal:

• Quantify quality and block regressions across PR/merge/release gates.

Use these modules:

• api/routes/evaluations.py
• evaluation/service.py
• evaluation/gates.py
• evaluation/regression.py

Typical flow:

1. Create run for gate type (G-PR, G-MRG, G-REL, G-MON).
2. Submit task results and quality metadata.
3. Compute metrics and gate verdict.
4. Save baseline for future comparison.
5. Triage/resolve regression records.

Best fit:

• Teams with golden-task style quality gates.

5. Memory-Backed Agent Sessions

Goal:

• Use retrieval context from long-term memory and session observations.

Use these modules:

• memory/long_term.py
• memory/rag.py
• memory/hybrid.py
• memory/progressive_recall.py
• api/routes/memory_search.py

Typical flow:

1. Store observations and embeddings.
2. Query memory via progressive recall levels (index, timeline, full).
3. Use RAG output as augmented prompt context.
4. Summarize sessions for long-horizon context compression.

Best fit:

• Research agents and recurring task assistants.

6. Continuous Improvement Operations

Goal:

• Track research intake, lessons learned, checklist completion, and roadmap refresh.

Use these modules:

• api/routes/improvements.py
• improvement/service.py
• improvement/models.py

Typical flow:

1. Submit and triage research intake.
2. Record lessons and action items.
3. Track checklist completion by period.
4. Capture metric snapshots and trends.
5. Record roadmap refresh outcomes.

Best fit:

• Teams running explicit continuous-improvement loops.

7. Multi-Channel Webhook Intake

Goal:

• Accept Telegram/Discord/Slack/WhatsApp webhook events into AGENT-33.

Use these modules:

• api/routes/webhooks.py
• messaging/*.py

Typical flow:

1. Instantiate adapter(s) in-process.
2. Register adapter(s) using register_adapter(platform, adapter).
3. Receive provider webhook callbacks via /v1/webhooks/*.
4. Enqueue platform events for downstream handling.

Best fit:

• Environments with external chat/event integrations.

Constraint:

• Adapters are not auto-registered in main.py; explicit bootstrap is required.

8. Prompt and Rollout Optimization (Experimental)

Goal:

• Collect rollouts and iterate prompts via optimization algorithms.

Use these modules:

• training/runner.py
• training/optimizer.py
• training/scheduler.py
• training/store.py

Typical flow:

1. Record rollouts and rewards.
2. Run optimizer across historical rollouts.
3. Persist prompt versions and metrics.
4. Revert to earlier versions when needed.

Constraint:

• API routes exist (/v1/training/*), but full runtime wiring (training_runner, agent_optimizer) is partial by default.