AI agent permissions and approval rules: Who or what an agent can access, and which actions need approval
Agent governance means deciding who or what an agent can access and which actions need approval. I explain the permission checks, approvals, tests, and records used to enforce those decisions. Related project pages describe their implementations and limits.
What governance means here
I start by listing the data and tools an agent needs, what it may do, and which actions a person must approve. Then I check how the system enforces those choices and what happens when a run fails. These are the checks I document:
- I check proposed actions against explicit rules before a provider call.
- I use acceptance tests to check the requested change and detect regressions.
- I run a fixed evaluation suite in continuous integration and enforce its release threshold.
- I connect requests, tool calls, and results with a shared trace ID.
- I check the remaining budget before issuing a provider call.
- I pause actions that require approval until a named person approves the exact payload.
- I document reversal procedures and identify changes that cannot be reversed.
The patterns
Each pattern page answers one question about running an agent safely. Start with the one closest to your problem.
Acceptance tests for agent work →
Executable checks test the task and detect regressions.
Answers: How do I check whether an agent completed its task?
Evaluation tests for releases →
A fixed evaluation suite checks release requirements.
Answers: How do evaluation tests check a release?
MCP tool permissions →
The server checks which tools a caller may use.
Answers: How do I restrict an MCP caller’s tool access?
Recording agent runs →
A shared trace ID connects the records of a run.
Answers: What should I record to investigate an agent failure?
Spending limits for agent runs →
Budget checks run before provider requests.
Answers: How do I check spending limits before an agent call?
Pausing for human approval →
The run waits for approval before the action executes.
Answers: How do I pause an agent for human approval?
Replaying and comparing agent traces →
Step-by-step comparison finds where recorded runs differ.
Answers: How do I compare recorded agent runs?
Recovering runs from checkpoints →
Stored checkpoints support recovery after interruption.
Answers: How do checkpoints help a run recover?
Reversing agent-made changes →
Each change needs a documented reversal procedure.
Answers: How do I reverse changes an agent made?
Handing a system over to a team →
A named owner practices running, debugging, and stopping the system.
Answers: How do I hand a system over to a team?
Related project examples
ClawKeeper →
ClawKeeper checks proposed Claude agent actions against explicit rules before a model call. Autonomous spend defaults to zero, and the database enforces an append-only audit log.
TechTide Harness Kit →
TechTide Harness Kit uses JSON Schema to declare skill permissions, risk tiers, network access, and approval requirements. Automated checks validate declarations and run prompt-injection fixtures.
CipherClaw →
CipherClaw analyzes recorded agent activity without model calls. It builds causal graphs, scores possible causes, and flags failure patterns.
Swarm 357 →
357 Claude agent roles across six business layers, with portable memory, enforced cost caps, a bash policy gate, and durable checkpoints.
FintheFinder →
FintheFinder pauses research runs for human approval and audits citations deterministically. Its repository includes contracts and test files.
TechTideAI2 →
TechTideAI2 separates agent decisions, application data, and audit output. It uses explicit state transitions, execution timeouts, and adversarial evaluation tests.
I explain the engineering here. If your product build needs agent permissions and approvals that work for a real customer, pilot, or launch, send me the brief.
Product implementation
Show me what exists, what is stuck, and what needs to work. I take ownership of the implementation, test the agreed user journey, and document the handoff.
Scope and price are agreed before work starts.
Logs over vibes. Production over theater.
© 2026 Alex Cinovoj · TechTide AI · Columbus, OH