A promotion conversation ends with 'not yet' but no specifics. You cannot tell what evidence would change the answer. An agent delivers a PR. You can only tell whether it meets the quality bar if you read every line. Guide resolves both. It grounds 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. The guidance does not come from impressions or generic advice. Verify agent work against the standard. You then review by exception. You do not review 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 generate --all
npx fit-guide --init
Guide needs gRPC service clients that
fit-codegen generates. It also needs a configuration
scaffold that fit-guide --init creates. After setup,
authenticate with Anthropic. Then process your standard data. See
the
Getting Started guide for engineers
for the full walkthrough.
After you configure it, 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