Agent Catalog
Reference agents that illustrate API usage and serve as baselines.
Available Agents
| Agent | Target Tasks | Highlights | File |
|---|---|---|---|
| ToyAgent | filesystem_hidden_config@1 | Cautious filesystem exploration, basic error handling | agents/toy_agent.py |
| NaiveLLMLoopAgent | filesystem_hidden_config@1 | Single-shot read/extract with one retry; minimal competence | agents/naive_llm_agent.py |
| RateLimitAgent | rate_limited_api@1 | Respects retry_after, retries transient failures, payload caching | agents/rate_limit_agent.py |
| ChainAgent | rate_limited_chain@1, deterministic_rate_service@1 | Handshake orchestration, payload resets, fatal/transient recovery | agents/chain_agent.py |
| CheaterSimAgent | All tasks (defense testing) | Intentionally probes sandbox to exfiltrate hidden state; expected to fail | agents/cheater_agent.py |
| OpsTriageAgent | log_alert_triage@1, config_drift_remediation@1, incident_recovery_chain@1 | Deterministic log/config triage, drift diffing, handoff chaining | agents/ops_triage_agent.py |
| LogStreamMonitorAgent | log_stream_monitor@1 | Patience + trigger detection across paginated log stream; record mode prototype | agents/log_stream_monitor_agent.py |
Usage
Each agent can be passed to the runner via the --agent flag. Agents are Python classes that implement the required reset/observe/act interface. The CheaterSimAgent is provided specifically to verify sandbox defenses and is expected to fail every run.