メニュー
Build or repair a deterministic Harbor task environment with explicit resources, network policy, health checks, and safe solver-visible boundaries.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Implement a legitimate Harbor Oracle solution and validate its behavior against an independent verifier without weakening the task contract.
Red-team a Harbor verifier for answer leakage, reward hacking, shared-container tampering, artifact manipulation, and semantic under-testing.
Calibrate a frozen tb-hard task using Harbor Oracle and real-agent rollouts, then distinguish authentic difficulty from ambiguity, flakiness, leakage, or infrastructure failure.
Create or evolve a canonical Harbor task inside a stable tb-hard case directory. Use after the design brief names observable acceptance criteria and at least three difficulty dimensions.
Audit a calibrated Harbor task for deterministic purchaser export, leakage, evidence completeness, and release readiness.
Research Harbor and Terminal-Bench 2 patterns to design an original tb-hard task. Use before approving a design brief or when checking whether a candidate is too shallow or derivative.
| name | harbor-environment-design |
| description | Build or repair a deterministic Harbor task environment with explicit resources, network policy, health checks, and safe solver-visible boundaries. |
| argument-hint | <case_id> |
| user-invocable | true |
| context | fork |
| agent | harbor-env-engineer |
| model | sonnet |
| effort | high |
| allowed-tools | Read Write Edit Glob Grep Bash |
Target case: $ARGUMENTS
Read the approved private design brief and preserve the public task contract. Modify only cases/<case_id>/task/environment/ and environment-related portions of task/task.toml.
Build from deterministic assets, pin data generation, declare network_mode, and keep the initial state genuinely unsolved. Use Compose only when multi-service behavior is part of the intended capability. Add health checks and sidecar artifact strategy when applicable.
Never copy solution, tests, private notes, evidence, skills, subagents, Git metadata, credentials, or caches into the solver image. Run make validate CASE=<case_id> after edits and return build commands, baseline-state evidence, resource/network decisions, and any recommendation for separate verification.
