Skip to main content
Run a multi-agent investigation. Agents take turns appending findings, evidence, and analysis to a shared findings document until quorum is reached. 
entire investigate [SEED_DOC] [flags]
entire investigate SUBCOMMAND
entire investigate is experimental and appears under labs while the workflow is still being refined.

​
Inputs

Use one of these inputs:
InputDescription
SEED_DOCPath to a starting findings file
--issue-link URLGitHub issue or pull request URL resolved through gh
When neither input is supplied and the multi-agent picker opens, the picker collects an investigation prompt that becomes the topic for the run.

​
Flags

FlagDescription
--agents AGENTSOverride configured agents with a comma-separated list
--max-turns NPer-agent turn budget. Defaults to 2.
--quorum NApprovals needed to terminate. 0 means all agents.
--continue RUN_IDResume an existing run
--editRe-open the investigate config picker
--findingsBrowse local investigation manifests
--allow-untrusted-seedRequired for non-interactive --issue-link investigations because the seed is attacker-influenced GitHub content

​
Subcommands

SubcommandDescription
show RUN_IDPrint a saved investigation’s summary and findings
fix RUN_IDLaunch a coding agent with the run’s findings as grounded context
clean RUN_IDDelete a saved investigation
clean --allDelete all saved investigation artifacts

​
Examples

entire investigate leaderboard-route-investigation.md --max-turns 1
entire investigate --findings
entire investigate show RUN_ID
entire investigate fix RUN_ID
entire investigate clean RUN_ID

​
Continue with

Investigate Why Code Exists

Walk through a complete provenance investigation.

labs

See experimental Entire workflows.