ShulkwiSEC/bb-huge / skills/curated/colang-gen
colang-gen
VerifiedGenerates NeMo Guardrails Colang (.co) files and YAML config blocks from a plain-language description of a chatbot's purpose, allowed behaviors, and constraints. Use this skill whenever a user wants to build guardrails for a chatbot, define allowed intents for an LLM, create an AI firewall with NeMo Guardrails, generate Colang flow definitions, or configure a semantic allow-list for a bot. Trigger this skill even when the user just describes what their bot should and shouldn't do — generating the Colang and YAML is almost always what they need next.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 7.4 KB |
| evals/evals.json | config | 5.5 KB |
evals/evals.json
View on GitHub{
"skill_name": "colang-gen",
"evals": [
{
"id": 1,
"prompt": "I'm building a support bot for a mobile banking app. It should handle account balance questions, card freezing/unfreezing, transaction disputes, and password resets. Nothing else. Formal tone.",
"expected_output": "A config.yaml using the standard security classifier template with allow_free_text: false, default_reply: false, output_scanning: true, colang_version 2.x. A main.co with intents covering at least: greeting, account balance, card freeze/unfreeze, transaction disputes, password reset, plus a strict fallback. Each user intent should have 5+ semantically varied example phrases.",
"files": [],
"assertions": [
{
"name": "config has allow_free_text false",
"description": "config.yaml must contain allow_free_text: false"
},
{
"name": "config has default_reply false",
"description": "config.yaml must contain default_reply: false"
},
{
"name": "config has output_scanning true",
"description": "config.yaml must contain output_scanning: true"
},
{
"name": "config has colang_version 2.x",
"description": "config.yaml must contain colang_version: \"2.x\""
},
{
"name": "config uses standard security classifier instructions",
"description": "config.yaml instructions block must contain the security classifier system prompt (not a generic bot description)"
},
{
"name": "colang has fallback flow",
"description": "main.co must contain a fallback flow with 'when true'"
},
{
"name": "colang covers all 4 required intents",
"description": "main.co must have flows for account balance, card freeze/unfreeze, transaction disputes, and password reset"
},
{
"name": "user intents have 5+ example phrases each",
"description": "Each define user block should have at least 5 example phrases"
},
{
"name": "no pricing or off-topic flows",
"description": "main.co should not contain flows for topics outside the stated scope"
}
]
},
{
"id": 2,
"prompt": "Need a bot for my bakery website. Customers ask about our menu, opening hours, custom cake orders, and allergens. Keep it friendly.",
"expected_output": "A config.yaml using the standard security classifier template. A main.co with intents for: greeting, menu questions, opening hours, custom cake orders, allergen info, plus fallback. Bot responses should be warm/friendly in tone. Each user intent has 5+ varied example phrases.",
"files": [],
"assertions": [
{
"name": "config has required security settings",
"description": "config.yaml must have allow_free_text: false, default_reply: false, output_scanning: true"
},
{
"name": "config uses standard security classifier instructions",
"description": "config.yaml instructions block must contain the security classifier system prompt"
},
{
"name": "colang covers all 4 required intents",
"description": "main.co must have flows for menu, opening hours, custom cake orders, and allergens"
},
{
"name": "colang has fallback flow",
"description": "main.co must contain a fallback flow with 'when true'"
},
{
"name": "user intents have 5+ example phrases each",
"description": "Each define user block should have at least 5 example phrases"
},
{
"name": "bot responses match friendly tone",
"description": "Bot response examples should use friendly/warm language, not formal/clinical"
}
]
},
{
"id": 3,
"prompt": "Building an onboarding assistant for a B2B SaaS product. It should help new users understand features, walk through setup steps, and escalate to human support if they're stuck. Block everything off-topic including pricing questions — sales handles those.",
"expected_output": "A config.yaml using the standard security classifier template. A main.co with intents for: feature overview, setup walkthrough (potentially multi-step), human escalation. Pricing must not appear as an allowed flow. Fallback must be present and strict.",
"files": [],
"assertions": [
{
"name": "config has required security settings",
"description": "config.yaml must have allow_free_text: false, default_reply: false, output_scanning: true"
},
{
"name": "config uses standard security classifier instructions",
"description": "config.yaml instructions block must contain the security classifier system prompt"
},
{
"name": "colang has human escalation flow",
"description": "main.co must include a flow for escalating to human support"
},
{
"name": "pricing is not an allowed intent",
"description": "main.co must not contain a flow that handles pricing questions"
},
{
"name": "colang has fallback flow",
"description": "main.co must contain a fallback flow with 'when true'"
},
{
"name": "user intents have 5+ example phrases each",
"description": "Each define user block should have at least 5 example phrases"
}
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |