AI Agent · Reviewed 2026-06-07

Mercury Agent by CosmicStack Labs

FADING · 40/100

Mercury Agent from CosmicStack Labs — a GitHub-hosted agent project where the organisational branding outpaces the visible product surface.

Visit Mercury Agent by CosmicStack Labs →

Mercury Agent from CosmicStack Labs has the hallmarks of a deliberate product attempt (organisational GitHub account, compound product name suggesting versioning intent) but the public surface at review time consists solely of the GitHub repository. The 'Mercury' naming in the agent space is somewhat crowded (Mercury is also a frontend build tool and several other OSS projects), but the 'CosmicStack Labs' organisation suggests an attempt to build a multi-product agent toolkit. Without a product page, documentation, or demo, evaluation is limited to what can be read from the repository directly. Assign to the re-review queue when a product site appears.

Why FADING

FADING (40) — GitHub-only surface from an organisational account. The branding suggests deliberate product intent but the public evaluation surface is not yet present.

What it does well

What it fails at

Red flags

Best for

  • Technical evaluators willing to read source code

Not recommended for

  • Production AI agent deployments
  • Buyers needing documented, stable tooling

Related agents

Agent relevance

Behavioral-testable

No documented integration path. Source-level inspection required.

Agent-friendly score: 2/10

Evidence

scorecard.json · registry · methodology

Verdict by Hlido Editor · Method: public-surface-tier-1+editorial-narrative-v2 · Methodology version 2026.05 · Next review due 2026-09-07