A promotion conversation ends with 'not yet' but no specifics — unclear what evidence would change the answer. An agent delivers a PR and the only way to tell whether it meets the quality bar is reading every line. Guide resolves both by grounding career advice and output review in the organization's actual engineering standard.
What becomes possible
For Empowered Engineers
Get guidance and evidence grounded in your organization's standard, not impressions or generic advice. Verify agent work against the standard so you review by exception, not by default.
- Gap analysis between current level and target, with specific skill recommendations
- Skill assessment that interprets engineering activity against standard markers
- Onboarding orientation grounded in actual team expectations
- Output review that checks deliverables against the organizational quality bar
Getting Started
npm install @forwardimpact/guide
npx fit-codegen --all
npx fit-guide --init
Guide needs gRPC service clients generated by
fit-codegen and a configuration scaffold created by
fit-guide --init. After setup, authenticate with
Anthropic and process your standard data — see the
Getting Started guide for engineers
for the full walkthrough.
Once configured, use Guide from any of three surfaces:
npx fit-guide # CLI (Claude Agent SDK)
# Or connect via Claude Code / Claude Chat using the MCP endpoint