Skip to content

Slash Commands Reference

agents-fleet registers 69 built-in commands plus 11 auto-discovered workflow commands (synthesized at startup from bundled workflow frontmatter). Project/user workflows may add more. Built-in source of truth: src/commands/registry.ts; workflow commands: src/skills/bundled/workflows/**. For workflow YAML see Workflows; for crews see Crews.

Generated for agents-fleet v0.52.0 · 69 built-in + 11 workflow commands.


UI

CommandAliasesDescription
/helpShow available commands
/docs/docBrowse and read the shipped fleet documentation in chat
/clear/clsClear the screen
/versionShow version information
/exit/quit /qExit the CLI
/restartRestart agents-fleet in a new PowerShell window with the SAME flags + cwd + env (Windows-only in v1). The new window inherits the current shell's env, preserving FLEET_HOME, AGENTS_FLEET_TELEGRAM_TOKEN, and security flags (--yolo, --acknowledge-*) exactly.
/copyCopy last assistant message to clipboard
/scroll/more /lessView conversation history in a scrollable pager
/historyShow conversation history (page by page)
/compactCompress conversation context
/viewSwitch chip-strip view, list views, or enter strip-focus mode
/nameSet or view session name

Sessions

CommandAliasesDescription
/sessionsInteractive session picker
/resumeResume a previous session
/configShow or update configuration
/costShow accumulated token costs
/tokens/usage /ctxShow token usage and context window status
/stats/statusShow session statistics
/doctor/diagRun diagnostic checks
/diagnoseLive runtime diagnostics for the coordinator (separate from /doctor which checks env)
/stall-historyShow recent coordinator stall + recovery events from the intel DB (T-214)
/model/mShow or switch the current model
/modelsInteractive model picker
/context-tier/context-windowView or set the context-window caps (coordinator + worker; default 200k / long_context 1M)
/effortShow or change reasoning effort level
/planEnter or exit plan mode

Fleet

CommandAliasesDescription
/workersShow active workers with progress
/stopStop a specific worker
/sendSend a message to a specific agent
/broadcastBroadcast a message to all team members
/fleet-log/log /activityShow agent activity log (or tail the log file)
/pause-dispatch/pausePause automatic task dispatch to workers
/resume-dispatch/unpauseResume automatic task dispatch to workers
/topologyView or set the messaging topology (hub, mesh, silent)
/teamTeam management (coordinator-managed; humans usually use /broadcast or /crew)
/team-repo/metaManage the shared team meta-repo (init/sync/push/promote/status)

Tasks

CommandAliasesDescription
/tasksShow task breakdown with dependency graph
/dod/definition-of-doneShow Definition of Done status and verify items
/start/go /runStart fleet execution of the task plan (auto-loops every 30m)
/loopAuto-execute on an interval until all tasks are complete (alias of /loop-target)
/loopsList and manage active loops
/loop-target/auto /executeAuto-execute on an interval until all tasks are complete

Workflows

CommandAliasesDescription
/new-project/kickoff /projectStart a guided project planning session (7 interactive phases)
/brainstorm/brain /bsAdversarial multi-agent pre-feature brainstorming. Spawns 3 inquisitor explorers to grill the user with questions until consensus.
/feature/feat /epicPlan a feature on an existing project, workflow-driven (brainstorm → requirements → research → design → validation → DoD → tasks)
/compete/race /bakeoffRun adversarial implementation competition (parallel builds in git worktrees)
/doPlan and execute a task with DoD tracking (research → plan → execute)
/fast-do/quickQuick execution — skip planning, up to 3 workers
/workflowsManage suspended workflow interrupts (list / status / resume / abort / release)
/approveResume the suspended workflow in this cwd with decision=continue
/rejectResume the suspended workflow in this cwd with decision=abort
/wfWorkflow observability + scaffolding — list runs, drilldown, scaffold autopilot, reload registry

Knowledge

CommandAliasesDescription
/skillsList available roles, atomic skills, and workflows (or migrate v1)
/skill/skCreate or list skills
/skills-migrateMigrate v1 .md skill/crew files to v2 (.role.md / .crew.md)
/crewsList available crews (v2 roster + workflow)
/crewActivate, deactivate, create, rollback, or show current crew
/learn/intelligenceReview fleet intelligence learnings, suggestions, and memory
/experiment/expDefine, run, and report A/B experiments over skills/roles/crews
/fleet-intel/intelView fleet intelligence insights from past sessions
/mcpManage MCP servers
/context-status/init-statusShow project context status (files generated by /init). Alias: /init-status (deprecated, removal next major)
/context-scopes/init-scopesList available context scopes from .fleet/context/scopes.json. Alias: /init-scopes (deprecated, removal next major)
/context-load/init-loadLoad a context scope into the current conversation. Alias: /init-load (deprecated, removal next major)

Git

CommandAliasesDescription
/diffShow git diff of current changes
/commitCreate a git commit (delegates to coordinator)
/worktree/wtManage git worktrees for worker isolation

Telegram

CommandAliasesDescription
/bot-set-tokenPersist Telegram bot token to ~/.fleet/telegram.json (mode 0600). The AGENTS_FLEET_TELEGRAM_TOKEN env var always wins over this file.
/bot-clear-tokenRemove Telegram bot token from ~/.fleet/telegram.json. Idempotent. Does not affect the AGENTS_FLEET_TELEGRAM_TOKEN env var.

Workflow Commands (auto-discovered)

CommandAliasesDescription
/bug-autopilot/bugfix-autopilotEnd-to-end bug-fix autopilot: reproduce → fix → regression-test → code-review loop → open PR → wait for human approval → iterate on feedback → merge.
/code-review/review /crRun comprehensive multi-agent code review
/code-review-autopilot/cr-auto /review-loopIterative code-review auto-pilot: review → triage → fix → validate, looping until convergence; optional PR handoff via --ship. Pass --scope=full to audit the whole codebase instead of a diff.
/fixer-loopAuto-pilot loop that reviews, triages, fixes, and verifies. Iterates until convergence or budget exhausted.
/full-review/cr-full /full-crFull-codebase report-only review (no fixes): audits the ENTIRE tree via the empty-tree..HEAD range with 5 parallel reviewers and writes CODE_REVIEW.md. Nothing is modified.
/github-issue-worker-autopilotGitHub-issues-driven autopilot: triage open issues, fix, validate with tests, generate proof artifacts, and ship as a draft PR.
/init/investigateWorkflow-driven /init — parallel investigators + critic + synthesizer
/research/investigate-topic /deep-diveAdversarial deep research on a topic (parallel explorers + critic)
/security-autopilot/sec-audit /security-auditDefensive application-security audit autopilot: recon → hunt (per-class fan-out) → triage → PoC → fix → verify → report. Static analysis + local sandboxed PoCs only; never touches live/third-party systems. Pass --ship to open a PR with the fixes; default writes SUMMARY.md + report.md only.
/security-binary-autopilot/binary-autopilot /sec-binaryBinary-analysis autopilot: target recon → static RE + patch-diff → coverage-guided fuzzing in ISOLATED-DISPOSABLE Azure VMs → crash triage + exploitability + minimal impact PoC → MSRC-format report → disclosure package. Authorized-scope only; isolated/disposable VMs; coordinated disclosure via MSRC only; no weaponization.
/security-disclosure-autopilot/sec-disclose /msrc-autopilotCoordinated-disclosure autopilot: recon → scan → triage → local PoC → ISOLATED-DISPOSABLE Azure-VM detonation + evidence capture → MSRC-format report → optional fix → disclosure package. Authorized-scope only; MSRC coordinated disclosure only; isolated/disposable VMs (no prod/third-party, auto-teardown); evidence integrity (hash+timestamp).