DevHunt Overview and Ethics

What the fixed source does

DevHunt collects GitHub activity for a requested login, stores the records locally, and builds descriptive summaries from them. It can also prepare an AI-written narrative from the resulting profile—or just save the prompt for your review.

All of this runs through authenticated GraphQL queries. Those queries follow whatever the token can see, with no public-only filter. A sufficiently privileged token can pull in private or internal-visible repository and pull-request data.

Three-stage workflow

The source exposes three commands:

1. scan — collects inputs from GitHub and writes raw JSONL files.
2. report — reads the raw files, derives a local profile, and generates charts.
3. narrate — turns the profile into prompts, or sends prompt material to a configured provider.

Each stage reads the output of the previous one. You can stop at any stage—for example, running only report and narrate with synthetic data, without ever scanning a real account.

Ethics and interpretation

Just because data is visible to the supplied token does not make it context-free. Repository history can be incomplete, shared, automated, private, or unrelated to employment.

The metrics in this revision are descriptive heuristics defined by the project. They are not validated measures of skill, personality, intelligence, employability, or objective ability. Treat them as pattern summaries with explicit sample sizes, not scores.

Safer use

Use the output to examine source-visible patterns and the sample sizes behind them. When a value is unknown, keep it unknown. Retain the underlying evidence, avoid ranking people, and remove per-user records when their purpose ends.

Continue

Head to Quick Start for a hands-on walkthrough, Metrics for exact formulas and unknown-value behavior, and Data Handling before collecting another person's token-visible metadata.