Audit Tools: Reviewing Agent Decisions

Updated May 22, 2026 · 1 min read

List, inspect, and query agent decision logs with the MCP audit tools, plus local runtime commands that skip the MCP round-trip.

Three MCP audit tools track agent actions:

audit_get_recent_actions

Lists recent mutation actions on a strategy with the AI's reasoning for each.

scope: Recent mutation actions across a strategy (open/close/edit/cancel) with the AI's reasoning per action.

audit_get_strategy_history

Lists parameter changes to a strategy itself, such as TP/SL, multiplier, and pause updates, with reasoning.

scope: Parameter mutations on the strategy itself (TP/SL changes, multiplier updates, pauses) with reasoning.

audit_query

Queries audit history with combined filters for action type, time range, strategy, and asset.

scope: General filter API. Combine action type, time range, strategy id, and asset.

On the runtime side, decisions are also cached locally:

openclaw senpi action list
openclaw senpi action inspect <id>
openclaw senpi action history
openclaw senpi action decisions

These give you per-runtime decision logs without an MCP round-trip. For deeper diagnosis when DSL state looks wrong, see DSL troubleshooting and state validation.

Skills source

Browse open-source Senpi skills on GitHub.

View on GitHub

supportguideSkills, Audit & Activity

Senpi Skills Library

Open-source Senpi trading skills for scanning and momentum: list them, use the runtime for exits, and migrate legacy Python via the playbook.

supportguideSkills, Audit & Activity

Share Your Wins Feed

How get_share_your_wins surfaces closed winning trades, the fields it returns, and why a trade might not appear in the feed.