raven
01Plugin Claude Code
Plugin marketplace listing 1 plugin: raven.
Plugin Claude Code
Plugin marketplace listing 1 plugin: raven.
Plugin Claude Code
AI codes fast. Raven enforces Discipline — Strategic Thinking, Scalable Structure, Security at Source. A plan attacked from three angles before any edit, approved by you; every prompt routed to the right one of 61 specialists, never silently; and local guards that block secrets, vulnerable libraries, and edits until…
Hook Claude Code
Runs when you submit a prompt, before the agent sees it, executing triage-router.py, architect-router.py, model-router.py, cve-prompt-guard.py and push-approve.py via python3 with --hook (5 commands). From giggsoinc/raven.
Hook Claude Code
Runs before the agent uses a tool for Write, Edit, MultiEdit, NotebookEdit and Bash tool calls, executing route-gate.py and push-gate.py via python3 (2 commands). From giggsoinc/raven.
Hook Claude Code
Runs when a session starts, executing session-start.py, ide-boot.py, model-router.py and run_costs.py via python3 with --session-start (5 commands). From giggsoinc/raven.
Hook Claude Code
Runs after a tool call finishes for Write, Edit and MultiEdit tool calls, executing db-guard.py and secret-scan.py via python3 with --changed-files-only (2 commands). From giggsoinc/raven.
Hook Claude Code
Runs when the agent finishes a response, executing token-meter-write.py, token-guard.py, dashboard.py, obsidian-log.py, knowledge-extract.py, session-gate.py and code-xray.py via python3 with --html, --current-project, --if-stale, --quiet, --delta and --session (7 commands). From giggsoinc/raven.
Settings file Claude Code
Raven v5.0 — Automated governance gate via Andie.
Skill Claude CodeCodex
Toggle the Raven model router. /router enables advisory Haiku-subagent delegation for simple prompts, /router off disables it, /router status shows the current mode. The primary session model is never changed — that is set only via /model.
Cursor rule Cursor
Raven router and cost must run every user turn.
Instructions file GitHub Copilot
Instructions for giggsoinc/raven, covering raven — github copilot / vs code and first load (same ux every ide).
MCP server Claude CodeCodexCursor +2
MCP server "raven" as configured in giggsoinc/raven. Runs ${workspaceFolder}/mcp/server.py with python3.
Instructions file CodexOpenCode
Instructions for giggsoinc/raven, covering agents.md — grok / codex / cursor boot and first load (same ux every ide).
Instructions file
Instructions for giggsoinc/raven, covering claude.md — raven discipline engine v5.5.6, first load (memory), step 0 — model toast (before andie, before tools), step 0.5 — intent (plan / debug / direct) and step 1 — gate check (default = raven, opt-out = lucky).
Instructions file Gemini CLI
Instructions for giggsoinc/raven: Fallback: python3 "${CLAUDEPLUGINROOT}/scripts/ops/raven-first.py" --boot.
Agent
Use PROACTIVELY when new code files are created. During coding — advises only, never blocks. At git commit — missing architecture doc becomes a hard block via pre-commit hook after 24 hours grace period. Ensures .raven/architecture.md exists and is versioned.
Agent
Use PROACTIVELY at session start and end to manage persistent memory. At start: run ide-boot.py; if load=1 Read CARD.md. At end: obsidian-log writes vault + CARD. Canonical vault is /RavenVault (humans). Agent boot is the card.
Agent
Database discipline enforcer. Fires on PostEdit for any source file. Detects inline SQL in non-SQL files, missing ERDs, and broken migration numbering. Spawned as a sub-agent by db-specialist pre-flight. Also runs independently on every file save via PostEdit hook.
Agent
Use PROACTIVELY when any skill attempts to read sensitive files or perform restricted actions. Watches for unexpected reads of manifest.secrets.json, .env, SSH keys, or modifications to settings.json or manifest.json by non-approved processes. Hard blocks and alerts.
Agent
Use PROACTIVELY before any coding action, file creation, or tool use. Verifies .raven/manifest.json exists, is valid JSON, and has all required fields. Hard stops if manifest missing — this is the only agent that hard blocks during coding. All other agents advise only.
Agent
Odoo discipline enforcer. Spawned by odoo-specialist pre-flight and fires on PostEdit for .py and .xml files. Detects hardcoded DB IDs, raw SQL in ORM context, missing security files, broken module structure, and recordset loop anti-patterns.
Agent
Salesforce discipline enforcer. Spawned by salesforce-specialist pre-flight and fires on PostEdit for .cls, .trigger, .js, .html files. Detects hardcoded IDs, SOQL inside loops, DML inside loops, missing bulk patterns, and non-DX file structure.
Agent
Use PROACTIVELY when any skill attempts to read sensitive files or perform restricted actions. Watches for unexpected reads of manifest.secrets.json, .env, SSH keys, or modifications to settings.json by non-approved processes. Hard blocks and alerts on any violation.
Agent
Use PROACTIVELY when any import or library is detected in code. During coding — warns and starts approval flow but does not block. At git commit — unapproved libraries are hard blocked via pre-commit hook until approval flow completes. Polars preferred over Pandas.