OpenCROW Utility Toolbox

Use this skill when the blocker is shell ergonomics or structured-data processing rather than a challenge-specific exploit primitive. It covers the “glue” layer that makes larger CTF workflows faster: jq, yq, xxd, tmux, screen, ripgrep, fzf, opencrow-autosetup, and opencrow-exploit.

Quick Start

Start the MCP server from the installed CLI:

opencrow-utility-mcp

Verify the mapped stack:

python ~/.codex/skills/opencrow-utility-toolbox/scripts/verify_toolkit.py

Workflow

1. Start with the MCP server and call toolbox_info, toolbox_verify, and toolbox_capabilities.
2. Use utility_search first when a workspace is large and you need to narrow the problem before opening files manually.
3. Use utility_json_query or utility_yaml_query when configs, API responses, or challenge metadata need slicing before deeper analysis.
4. Use utility_hexdump when you need a fast bounded hex view of a file region.
5. Use tmux or screen when the task benefits from persistent panes or background sessions.
6. Use opencrow-autosetup to seed a standard challenge workspace with OpenCROW reconnaissance artifacts, maintain HANDOFF.md, and launch a reconnaissance-only Codex pass that selects the final category-specific AGENTS.md.
7. Use opencrow-exploit to launch the follow-up Codex solve pass from the AGENTS.md and HANDOFF.md artifacts created by autosetup.

Resources

• opencrow-utility-mcp: stdio MCP server for typed workspace search, jq/yq queries, and xxd hexdumps.
• scripts/verify_toolkit.py: confirm that the mapped workflow helpers are installed.
• references/tooling.md: quick guidance for common shell utility choices.