Traction scoreGitHub stars can be faked, so popularity alone can be misleading. Traction Score looks for broader signs of recent attention, adoption, and active maintenance.
This kit provides a systematic approach to developing AI agents and commands, grounded in Anthropic's best practices. It helps automation builders move from raw prompts to validated skills through rigorous evaluation and structured specification.
The collection includes tools for root cause analysis, task drafting, and knowledge base reconciliation, ensuring that agent behavior remains aligned with the current state of a repository.
WHO IT'S FOR
AI agent / automation builders
developing complex skills using Anthropic best practices
Technical leads / architects
performing root cause analysis and technical specification
Compatible AgentsThe repository documents support for these agents. The skills may also work with other agents that can load SKILL.md files, but they may need some setup or small changes.
Installs: A user-selected set of Context Engineering Kit skills and agents; this option selects create-ideas, why, and status. · Cursor, Antigravity, Codex, OpenCode
Before you start
`npx` available
Run this command in a terminal. When the installer asks which skills and agents to install, select `create-ideas`, `why`, and `status`.
A question, problem, or opportunity to generate ideas for; replace `[question or problem]` with it.
Use the create-ideas skill to generate six genuinely different responses to this question: [question or problem]. Include text and a numeric probability for each response, with three likely answers above 0.80 and three diverse tail ideas below 0.10.
A concrete problem or symptom; replace `[issue description]` with it.
Use the why skill to analyze this issue with the default depth of five: [issue description]. State the problem clearly, document each why, validate the causal chain backward, explore branches if multiple causes emerge, and propose solutions that address the root causes.
A project directory containing an existing `.fpf/` knowledge base, or a directory to check for initialization; replace `[project directory]` with its path.
Use the status skill in [project directory] to inspect the existing `.fpf/` knowledge base. Report whether it is initialized, its current phase, hypothesis counts by layer, evidence freshness, decision count, and any warnings. Do not reset or modify the knowledge base.