I Built a Hermes Agent to Tell Me Which Hackathons to Enter. It Told Me to Enter This One.

Description and use

Vigil Crest is a Telegram‑based Hermes Agent that automatically scans the DEV challenge feed, evaluates each active hackathon on Time, Learning, Stack fit, and Timing, and returns a confidence‑rated verdict (enter/skip/maybe), saving time and improving participation decisions.

Problem

Choosing the right hackathon to enter is time‑consuming and risky; the DEV challenge feed is unsorted and lacks fit signals, leading to wasted effort.

Outcome

Vigil Crest triages active challenges with confidence‑rated verdicts, reducing wasted effort and improving hackathon participation decisions.

Workflow

• 1. Telegram bot sends a bi‑weekly reminder to check the challenge board.
• 2. User replies to trigger the agent.
• 3. Agent uses Playwright to browse dev.to/challenges and scrape active challenges.
• 4. For each challenge, the agent evaluates fit against a persona‑defined stack file and generates four fit lines and a verdict.
• 5. Agent sends the verdicts back to the user via Telegram.
• 6. User reviews the verdicts and decides whether to enter the challenge.

Evidence

Vigil Crest does not pretend to know me better than it does. Early on it labels its verdicts as first impressions and tells me what it is unsure about.

Source

It starts with the nudge. Twice a week Vigil Crest sends a short reminder to check the board.

Source

The agent browses the live challenge feed and reports back what it found.

Source