One-shot short observation window for running child agents. Returns only natural-language message events, and optionally normalized tool_call events, observed during this call; not a history API, not gapless streaming, and not stdout/stderr tailing. In v1, message extraction is supported for Code...
AI agents call peek to retrieve information from Agent Bridge without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
The tool reads/observes output from running subprocesses without modifying or controlling them. It retrieves message and tool_call events from child agents. Severity is medium because it can expose potentially sensitive output from AI agent sessions running arbitrary tasks, but the tool itself only reads data.
From the tool's definition One-shot short observation window for running child agents. Returns only natural-language message events... not a history API, not gapless streaming
Attacks that exploit this kind of access
One-shot short observation window for running child agents. Returns only natural-language message events, and optionally normalized tool_call events, observed during this call; not a history API, not gapless streaming, and not stdout/stderr tailing. In v1, message extraction is supported for Codex, Claude, OpenCode, Antigravity, and best-effort Forge Summary/Completed successfully lines. Forge tool calls are low-precision Execute/Finished markers and never include command output. Tool calls exclude raw tool output. It is categorised as a Read tool in the Agent Bridge MCP Server, which means it retrieves data without modifying state.
Register the Agent Bridge MCP server in PolicyLayer and add a rule for peek: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Agent Bridge. Nothing to install.
peek is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the peek rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for peek. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
peek is provided by the Agent Bridge MCP server (lailai258/agent-bridge-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →