src/simba/codex_skills/simba-onboard/SKILL.md
Analyze project markdown instruction files and generate consolidated SIMBA core instructions with markers. Use when Codex needs to onboard a repo by reading CLAUDE.md/AGENTS.md and .claude docs, then producing .claude/rules/CORE_INSTRUCTIONS.md (or configured filename), updating core reference blocks, and verifying markers.
npx skillsauth add mahmoudimus/simba simba-onboardInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Set up SIMBA markers for a project by analyzing existing markdown files and generating consolidated core instructions.
Follow these steps in order.
Run:
simba config get guardian.core_filename
Use the returned filename as ${CORE_FILE} for all following steps. Default is CORE_INSTRUCTIONS.md.
Collect onboarding inputs:
CLAUDE.md if it exists.AGENTS.md if it exists..md files under .claude/, excluding .claude/handoffs/ and .claude/notes/.simba markers audit.simba markers list.Report all files found, approximate sizes, and any existing markers or <!-- CORE --> tags.
Extract and categorize instructions from discovered files:
| Category | What to extract | SIMBA section |
| --- | --- | --- |
| Critical Constraints | Non-negotiable rules and security requirements | constraints |
| Build & Test | Build/test commands and CI steps | build_commands |
| Environment | Paths, ports, machines, deployment targets | environment |
| Code Style | Formatting and naming conventions | code_style |
| Workflow | Git/PR conventions and process rules | workflow |
| Agent Rules | Agent/subagent dispatch and context rules | agent_rules |
Rules:
Report a table of category, source files, and item counts.
.claude/rules/${CORE_FILE}Prepare a full draft with SIMBA markers:
<!-- BEGIN SIMBA:{name} --> and <!-- END SIMBA:{name} -->.SIMBA:core tight and critical only.---.SIMBA:core is re-injected repeatedly by guardian logic, so keep it concise.
Show the full proposed file first.
For each section include:
SIMBA:core that it is repeatedly re-injected.Request explicit user approval before writing files.
After approval:
.claude/rules/ exists..claude/rules/${CORE_FILE}.CLAUDE.md with a core reference block if missing:<!-- BEGIN SIMBA:core_ref -->
**Read `.claude/rules/${CORE_FILE}` for rules that apply to ALL contexts (main session + subagents).**
When dispatching subagents, inject the contents of that file into the prompt.
<!-- END SIMBA:core_ref -->
AGENTS.md similarly if it exists:<!-- BEGIN SIMBA:core_ref -->
**All agents must follow `.claude/rules/${CORE_FILE}` before executing.**
When dispatching write-capable agents, read `.claude/rules/${CORE_FILE}` and inject its contents into the dispatch prompt.
<!-- END SIMBA:core_ref -->
Run:
simba markers audit
simba markers list
Report final marker status and next recommended edits.
development
Index the current project for optimized search with QMD semantic search and fast file suggestions. Run this when entering a new codebase or after significant changes. Saves 60-80% tokens on exploration tasks.
tools
Show token economics comparing usage with turbo-search vs without. Demonstrates actual savings from search-first approach.
development
Enforce Simba's Codex lifecycle routine for coding tasks. Use when starting or finishing implementation work in a Simba-enabled repo to run `simba codex-status` at start, `simba codex-extract` when extraction is pending, and `simba codex-finalize` before final handoff.
tools
Save the current work session to persistent memory for future context. Summarizes accomplishments, tracks files modified, and stores learnings for cross-session continuity.