Software Development
Agent Tool Contract Reviewer
Agent Tool Contract Reviewer turns ai engineering material into a review report with a concrete agent tool-contract boundary check, tool prompt or evaluation failure mode, and contract agent implementation acceptance test. It is built for teams that need a usable draft artifact, not a broad summary with a familiar title. The skill asks for agent workflow, prompt, tool, or eval artifact, expected behavior and failure examples, and model, tool-call, and approval constraints up front, then organizes the output around agent, tool, contract, reviewer, and engineering. It is delivered with adapters for Claude Code, Codex, OpenClaw, and Antigravity so teams can use the same workflow across their preferred agent environment. Agent Tool Contract Reviewer is scoped to the AI Engineering workflow and is distinct from adjacent catalog skills by producing this specific deliverable rather than a broad summary.
One-Time Purchase
$19.99
Agent Tool Contract Reviewer — Sample Brief
Headline
Builds an agent-engineering review artifact for agent tool contract reviewer work, turning agent workflow, prompt, tool, or eval artifact, expected behavior and failure examples, and model, tool-call, and approval constraints into agent tool-contract boundary check, tool prompt or evaluation failure mode, and contract agent implementation acceptance test. In this example, Northbeam Analytics is reviewing an agent tool-call contract against expected schemas, failure cases, permission boundaries, and rerunnable eval evidence. The output is a concrete review report: source evidence, reviewer checks, and the owner decision that determines whether the work is ready.
Evidence Packet
| Input | Example | Status |
|---|---|---|
| agent workflow, prompt, tool, or eval artifact | Northbeam Analytics is reviewing an agent tool-call contract against expected schemas, failure cases, permission boundaries, and rerunnable eval evidence. | Provided |
| expected behavior and failure examples | Three representative examples from the current workflow | Reviewed |
| model, tool-call, and approval constraints | Approval threshold and final reviewer | Confirm |
| desired review, regression, or implementation target | AI Engineering owner and downstream audience | Ready |
Agent Evidence Model
| Field | What must be captured | Review consequence |
|---|---|---|
| Agent signal | Map tool name, expected input and output schema, permission boundary, failure examples, and approval gate separately. | Treat as draft until agent tool-contract boundary check is source-backed. |
| Agent Tool Contract Boundary Check | Tie prompt and tool-call behavior to eval evidence and contract tests rather than contract negotiation framing. | Escalate when agent evidence is missing or contradicted. |
| Tool Prompt Or Evaluation Failure Mode | Produce rerunnable acceptance tests for the tool contract before recommending approval. | Reviewer resolves contract agent implementation acceptance test before final use. |
Agent Acceptance Checks
| Check | Required evidence | Review action |
|---|---|---|
| Agent depth | Map tool name, expected input and output schema, permission boundary, failure examples, and approval gate separately. | Keep, revise, or ask a targeted follow-up |
| Tool boundary | Tie prompt and tool-call behavior to eval evidence and contract tests rather than contract negotiation framing. | Confirm this is not a renamed adjacent bundle skill |
| Contract handoff | Produce rerunnable acceptance tests for the tool contract before recommending approval. | Assign the owner decision before final use |
Agent Tool Contract Reviewer Work Map
Agent Tool Contract Reviewer focus areas
Draft Review Report
| Agent signal | Agent Tool Contract Boundary Check | Tool Prompt Or Evaluation Failure Mode | Contract Agent Implementation Acceptance Test | Tool decision |
|---|---|---|---|---|
| Tool contract | Agent input and output schema is explicit and versioned | tool prompt or evaluation failure mode | contract agent implementation acceptance test | Keep if covered |
| Failure case | agent tool-contract boundary check | Tool behavior is tied to an eval or replayable fixture | Agent evidence is missing or contradicted. | Revise until tested |
| Approval gate | Map tool name, expected input and output schema, permission boundary, failure examples, and approval gate separately. | Tie prompt and tool-call behavior to eval evidence and contract tests rather than contract negotiation framing. | Contract criterion names the reviewer and rerun command | Ready for review |
Use now
Agent Tool Contract Reviewer
Best when ai engineering work needs a concrete review report.
Hold back
Missing evidence
Best deferred when a key input would change the artifact.
Reviewer note
The output is structured so an operator can validate the draft, edit the artifact, and hand it to the next owner without re-reading the entire packet. The final answer should label unknowns plainly and avoid inventing metrics, policies, quotes, or commitments.
This sample illustrates the skill's output format. Companies referenced are fictional and used recurringly across sample outputs.
View full sample →
All sales final. No refunds on digital products.
Includes support for Claude Code, Codex, OpenClaw, and Google Antigravity in the same license.
Also in AI Engineering
Bundle price: $55. Compare this skill with the full workflow bundle or Pro access.
Best for
AI engineers reviewing an agent’s tool contracts and prompts — boundary checks, failure modes, and acceptance tests for tool-call behavior. Most useful when an agent calls real tools and a malformed contract or unhandled failure causes silent wrong actions.
Not ideal for
General application code review unrelated to agent behavior, and validating model quality, which needs eval datasets. It reviews the contract surface — it does not run the evals or guarantee the model behaves.
Included in this purchase
- Claude Code, Codex, OpenClaw, and Google Antigravity skill files.
- Setup guidance for the right adapter in your workspace.
- One-time license for the purchased skill version.
Setup
Plan for a short setup in the repository or workspace where the skill will run. Some coding familiarity helps for implementation-heavy outputs.
Related Skills
$19.99
One-time license
$19.99
One-time license
$19.99
One-time license
Future Updates
This purchase includes the current version of the skill. If you want future adapter updates — meaning compatibility and packaging updates as supported platforms evolve — plus new catalog additions included automatically, upgrade to Pro.