Building ai products
Skill VKK-00/agentic-project-lifecycle/plugins/agentic-project-lifecycle/skills/building-ai-products
Codex plugin for governing large software projects from discovery through release and operations.
npx -y skills add VKK-00/agentic-project-lifecycle --skill building-ai-productsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 16 days oldThe repository was created 16 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use when product behavior depends on LLMs or other models, prompts, retrieval, tool calling, AI evaluation, model routing, grounding, safety, latency, token cost, provider fallback, or model rollback.
The file declares its own license as Apache-2.0. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
1.8 KB, as published. Nobody here has run it
Building AI Products
Treat model behavior as a versioned, measured, failure-prone product dependency. Keep product policy outside opaque prompt text.
Bound model authority
- RULE-AI-01: State why AI is needed, which decisions remain deterministic, what the model may do autonomously, when human confirmation is required, and the fallback when the model is unavailable or uncertain.
- RULE-AI-02: Version a model/prompt contract with purpose, input schema, output schema, grounding sources, prohibited behavior, retry policy, fallback, and ownership.
Evaluate and operate
- RULE-AI-03: Create a representative held-out golden dataset and compare the candidate against a baseline using task-specific thresholds and guardrails. Record failures by category, not only one aggregate score.
- RULE-AI-04: Set latency, cost, context, routing, caching, and provider-failure budgets. Test fallback behavior and model deprecation before launch.
- RULE-AI-05: Threat-model prompt injection, data leakage, tool misuse, unsafe automation, unsupported claims, and sensitive-data handling. Require grounding or explicit uncertainty where correctness matters.
- RULE-AI-06: Do not widen exposure without versioned prompts/models, production sampling, regression evals, observability, a kill switch, and rollback to a known baseline.
Read product boundary, evaluation system, and operations and safety.