ceo-skill is designed to be portable across agent runtimes.
The skill is strongest when the runtime can provide one or more of the following:
- web search / browsing
- for market context, competitors, benchmarks, and recent events
- file reading
- to load references and examples
- optional code execution
- to run quantitative analysis with
scripts/analysis_tools.py
- to run quantitative analysis with
At minimum, the skill still works as a structured decision workflow if the runtime only supports text generation and file reading.
The experience improves significantly when the runtime can also:
- search the web for current facts
- run Python for EV / NPV / IRR / Monte Carlo
- access organization-specific context files or memos
This repository avoids over-coupling to a single platform. Most of the value is in:
- the decision flow in
SKILL.md - the reference documents in
references/ - the examples in
examples/ - the analysis helpers in
scripts/analysis_tools.py
- start by reading
SKILL.md - test with 3-5 real decision prompts
- add your own org-specific examples
- optionally wire the script helpers into your runtime tooling