Ross ROSS = Recommend OSS · open-source software intelligence for agents

rpamis/comet

Comet: agent skill harness for turning ideas into evaluated workflows observed · 2026-08-28

github.com/rpamis/comet · homepage · JavaScript · MIT (permissive) observed · 2026-08-28

Health v2 · maintenance only

80/100

  • Activity 99
  • Release rhythm 98
  • Longevity 7

Flags: young

How is this computed?

round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.

  • gap_med: 0.5
  • age_days: 111
  • days_rel: 12
  • days_push: 7
  • n_releases_24m: 47

Full methodology

Adoption not part of the score

2849 stars · 279 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Comet is an npm-distributed CLI and Skill platform that gives AI coding agents a resumable, phase-guarded workflow for long-running tasks, with Native and Classic spec-driven modes invoked via /comet. It also includes Comet Eval, a dual-agent evaluation harness that runs real tasks against any local Skill in isolated environments with rubric scoring and failure attribution.

Use cases

  • run resumable long-running coding tasks with an AI agent
  • evaluate AI agent skills on real tasks with automated scoring
  • turn team workflows into reusable agent skills
  • enforce spec-driven development phases in Claude Code or Codex
  • resume interrupted agent work across devices with zero context
  • benchmark coding agent workflows with pass@k metrics
  • compose existing skills like TDD and brainstorming into new workflows

When to choose

  • you use AI coding agents (Claude Code, Codex, Cursor, Gemini CLI) and need structured, recoverable workflows
  • you want measurable evidence that an agent skill actually works before publishing it
  • your team needs spec-driven development with phase guards and confirmation points
  • you want to package and distribute reusable skills across projects

When to avoid

  • you need a simple one-shot code generator without workflow overhead
  • you don't use any skill-capable AI coding platform
  • you require a fully stable release - the current version is beta and /comet-any is experimental
  • you cannot provide Docker, Python 3.11+, and model credentials for evaluation runs

Facets

cli-tool · maturity active

agent-framework workflow-automation llm-inference developer-tools testing cli developer-tools large-language-models cli cross-platform python skill-harness spec-driven-development phase-guards eval-harness coding-agents resumable-workflows skill-creator langsmith langfuse ai-agents automation command-line nodejs docker

9 sources

Member repositories

RepositoryRoleHealth v2
rpamis/cometmain80

For agents

markdown · JSON · MCP: product_card(name="rpamis/comet")

Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem