Slash commands - full reference

Navin ships built-in slash commands. Type / in any chat composer to open the command palette. Commands fall into three lifecycles:

Agent workflow - rewrites your message into a mission brief (with the right skills preloaded) and runs a normal agent turn.
Side channel - answers immediately without consuming an agent turn.
Turn control - manages the current chat/turn itself.

Most security / quality workflows are also available from the Code module Actions menu.

Agent workflows

Command	Title	Arguments	What it does
`/blueprint`	Plan mode	`[task]`	Designs an implementation plan before writing any code: constraints, options, chosen approach, step list. No code changes.
`/forge`	Build mode	`[task]`	Full autonomous build mode: plan, code, run, test, and iterate until done.
`/atlas`	Project atlas	`[init\|refresh\|query <question>]`	Builds or refreshes `.metadata/` (role, kind, and dependencies of every file) powering the Dev workbench Graph tab, and answers "where is X?" questions from the index instead of grepping.
`/inspect`	Code review	`[path\|diff\|scope]`	Reviews recent changes or a target for bugs, behavioral regressions, security issues, missing tests, and maintainability. Cites files and lines.
`/turbo`	Performance audit	`[path\|scope]`	Profiles and optimizes: hot paths, N+1 queries, blocking I/O, missing caches, bundle size, memory. Measures before recommending.
`/pulse`	Metrics & analytics	`[focus]`	Project health scoreboard: code size/complexity, dependency freshness, lint findings, coverage, TODO debt, KPIs.
`/board`	Task board	`[task <id>\|loop\|status\|plan <goal>]`	Works the shared project board: plan, take tasks, dispatch, and report.
`/goal`	Long-running goal	`<goal>`	Registers the request as a sustained goal the agent keeps pursuing across turns until completed or stopped.

Command	Title	Arguments	What it does
`/fortify`	Security review	`[path\|scope]`	Security audit: auth flows, input validation, injection surfaces, secrets handling, unsafe defaults, dependency risks. Findings rated critical → low.
`/probe`	Vulnerability scan	`[path\|scope]`	Hunts exploitable weaknesses: OWASP Top 10, hardcoded secrets, vulnerable dependency versions, SSRF/path traversal, prompt injection.
`/unmask`	Secrets scan	`[path\|scope]`	Leaked keys, tokens and passwords in code, config and git history.
`/lineage`	Supply-chain audit	`[path\|scope]`	Dependencies: CVEs, outdated packages, lockfile integrity, licenses.
`/xray`	Deep static analysis	`[path\|scope]`	SAST: traces user input to dangerous sinks across the code.
`/gatekeeper`	Access control audit	`[path\|scope]`	Authentication and authorization: sessions, tokens, roles, IDOR.
`/perimeter`	API & web surface	`[path\|scope]`	Exposed endpoints: authz, CORS, CSRF, SSRF, headers, rate limits.
`/bastion`	Infra & IaC audit	`[path\|scope]`	Docker, Kubernetes, Terraform, CI/CD and cloud configuration.
`/vault`	Data & privacy audit	`[path\|scope]`	PII handling, encryption, logging leaks, retention, GDPR gaps.

Command	Title	Arguments	What it does
`/recon`	Reconnaissance	`[path\|target]`	Maps the attack surface: routes, services, subdomains, tech fingerprint.
`/threatmap`	Threat model	`[path\|scope]`	Attack surface, trust boundaries and STRIDE threats with mitigations.
`/dast`	Dynamic testing	`[target\|scope]`	Runs the app and probes live: injection, auth, IDOR, SSRF, XSS/CSRF.
`/redteam`	Attack simulation	`[path\|scope]`	Chains weaknesses into realistic exploit paths with safe proof-of-concept.
`/pentest`	Full autonomous pentest	`[target\|scope]`	End-to-end cycle: recon, threat model, scan, exploit with PoC, validated report.
`/comply`	Compliance mapping	`[framework\|scope]`	Gap analysis vs OWASP ASVS, CIS, SOC 2, ISO 27001, PCI-DSS.
`/report`	Pentest report	`[format\|scope]`	Compiles validated findings into a shareable, compliance-ready report.

Command	Title	Arguments	What it does
`/studio`	Document studio	`[format + brief]`	Creates polished PowerPoint, Word, PDF, or Excel documents. See navin_contenant.
`/campaign`	Marketing studio	`[brief]`	Generates campaigns end to end: ad copy, social posts, product images, video scripts. See navin_marketing.
`/seo`	SEO studio	`[url\|topic]`	Full SEO: technical audits, keyword research, optimized content, competitor analysis. See navin_seo.
`/leads`	Leads & sales studio	`[icp\|company\|brief]`	Expert prospecting: company/people/job search, buying signals, scoring, outreach. See navin_leads.
`/team`	Team studio	`[mission\|org brief]`	Org design, role sheets, RACI, hiring, OKRs - and virtual AI teams via subagents. See navin_equipe.

All workflow commands accept an optional focus/target. Without one, the agent infers the most useful scope itself.

Command	Arguments	What it does
`/checkpoint`	`[save [note]\|list\|restore <name> [all\|chat\|code]\|delete <name>]`	Checkpoints are auto-saved before each prompt. Rewind the conversation, the code, or both. They complement git, not replace it.
`/pilot`	`[task]`	Model-per-task routing. Tasks: `search`, `plan`, `review`, `security`, `dev`, `fast`, `deep`, `docs`. Define presets with those names in Settings → Models; `/pilot review` then switches to the `review` preset. Without argument, shows the current mapping.
`/pack`	`[list\|install <git-url\|path>\|enable <name>\|disable <name>\|remove <name>]`	Manages plugin packs (skills + MCP servers). See Plugins.
`/model`	`[preset]`	Shows or switches the active model preset.
`/status`	-	Runtime, provider, and channel status (including web-search quota when available).
`/skill`	-	Lists all enabled skills available to the agent.
`/help`	-	Lists available slash commands.
`/history`	`[n]`	Prints the last N persisted conversation messages.
`/trigger`	`<name>`	Creates a named CLI trigger bound to this chat session.
`/pairing`	`[list\|approve <code>\|deny <code>\|revoke <user_id>]`	Manages DM pairing requests per channel.
`/restart`	-	Restarts the Navin process.

Command	Arguments	What it does
`/dream`	-	Manually triggers the two-phase memory consolidation.
`/dream-log`	-	Shows what the last Dream consolidation changed.
`/dream-restore`	-	Reverts memory to a previous Dream snapshot.
`/dream-prompt`	`[init]`	Customizes how Dream organizes this workspace's memory.
`/evaluator-prompt`	`[init]`	Customizes the heartbeat notification gate prompt.

Command	What it does
`/new`	Resets this chat and starts a fresh conversation (finalizes the active turn first).
`/stop`	Cancels the active agent turn for this chat.

Combine with the file target: after opening a file in the editor, the Actions menu can scope any command to that file (e.g. /inspect src/App.tsx).
Chain workflows: /blueprint first, review the plan, then /forge to execute it.
Pentest cycle: /recon → /threatmap → /probe or /dast → /redteam → /comply → /report (or /pentest to run the whole cycle).
Use /pilot deep before a hard investigation, /pilot security before an audit, /pilot fast for quick chores.
/checkpoint save before-refactor before risky operations, /checkpoint restore before-refactor code to roll back files only.
Full menu reference: Actions.