Agent skill

Comet

by rpamis in rpamis/comet

Comet — OpenSpec + Superpowers dual-star development workflow.

MITAuto-check passedAgent Workflows

Install Comet

skills CLI
$ npx skills add rpamis/comet --skill comet -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install rpamis/comet comet --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/rpamis/comet.git skills-src && mkdir -p .claude/skills && cp -r skills-src/eval/local/skills/benchmarks/039-release/comet-classic-039 .claude/skills/comet && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
comet
GitHub stars
3.2k
Token cost
~4.4k tokens
SKILL.md length
1,954 words
Files
17 (incl. scripts)
Skills in repo
52
Repo updated
First seen
Licence
MIT

At a glance

Comet — OpenSpec + Superpowers dual-star development workflow.

  • Works in 2 steps: Detect presets first; if hotfix/tweak… → When no preset matches, run openspec…
  • Agent Workflows work in your project
  • SKILL.md covers Decision Core, Subcommand Quick Reference and Reference Appendix
  • Runs Shell scripts from its folder

What it does

Comet is an agent skill from rpamis/comet. Comet — OpenSpec + Superpowers dual-star development workflow. Start with /comet for automatic phase detection and dispatch to subcommands. Five phases: open → design → build → verify → archive.

Its SKILL.md is about 4.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 19 other files, including scripts (for example `reference/auto-transition.md`, `reference/comet-yaml-fields.md` and `reference/context-recovery.md`).

It sits in Agent Workflows. The repository describes itself as: Comet: agent skill harness for turning ideas into evaluated workflows. The licence is MIT.

When your agent uses it

  • Agent Workflows work in your project

Example prompts

  • “/comet”

Requirements

  • A Bash shell

Workflow steps

2 steps, taken from the first numbered list in SKILL.md.

  1. Detect presets first; if hotfix/tweak matches, invoke the corresponding preset skill directly and do not enter the normal open branch
  2. When no preset matches, run openspec list --json to get all active changes

What it can do on your machine

Read from SKILL.md and the folder at commit 0fd42a0. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 7 files in scripts/ (Shell), which the agent can run.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Comet loads about 4.4k tokens when it runs. Until then it costs about 50 tokens; SKILL.md has 1,954 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~50
When it runs · the whole SKILL.md, loaded when a task matches
~4.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from rpamis/comet at commit 0fd42a0, republished under its MIT licence (© rpamis). 1,954 words, ~4,362 tokens.

Download SKILL.mdSave it as .claude/skills/comet/SKILL.md (or your agent's skills folder). This skill also uses 16 other files; get the full folder from GitHub.
name
comet
description
Comet — OpenSpec + Superpowers dual-star development workflow. Start with /comet for automatic phase detection and dispatch to subcommands. Five phases: open → design → build → verify → archive.

Comet — OpenSpec + Superpowers Dual-Star Development Workflow

OpenSpec and Superpowers orbit the same goal like a binary star system.

OpenSpec handles WHAT  — outline, proposal, spec lifecycle, archive
Superpowers handles HOW — technical design, planning, execution, closing

Core principle: brainstorming cannot be skipped. Every change must undergo deep design (except hotfix and tweak presets).


Decision Core

Agents need only read this section for decision-making. Refer to the Reference Appendix as needed.

Output Language Rule

Use the language of the user request that triggered this workflow as the default output language. When resuming an existing change with a clear dominant artifact language, preserve that language unless the user explicitly asks to switch.

Automatic Phase Detection

Step 0: Active Change Discovery and Intent Detection

  1. Detect presets first; if hotfix/tweak matches, invoke the corresponding preset skill directly and do not enter the normal open branch
  2. When no preset matches, run openspec list --json to get all active changes

Preset detection has highest priority:

  • User explicitly describes a bug fix / hotfix + meets hotfix conditions → directly invoke /comet-hotfix
  • User explicitly describes copy/config/docs/prompt small adjustment + meets tweak conditions → directly invoke /comet-tweak
  • No preset match → follow the table below
Active changesUser inputBehavior
Nonenon-preset input→ Invoke /comet-open
Exactly 1/comet <description>→ Ask: continue this change or create a new change
Multiple/comet <description>→ Ask: continue existing or create new; if continuing, list changes for selection
Exactly 1/comet with no description→ Auto-select, enter Step 1
Multiple/comet with no description→ List changes for user selection
<IMPORTANT>
When the user chooses "create a new change", **must invoke `/comet-open`**. Do not call `/opsx:new` directly.
`/comet-open` performs dual initialization: OpenSpec artifacts (created by internal `/opsx:new`) plus `.comet.yaml` state file.
Calling `/opsx:new` directly leaves `.comet.yaml` missing and breaks later phase detection.
</IMPORTANT>

Step 1: Read .comet.yaml state metadata

Prefer reading openspec/changes/<name>/.comet.yaml. If not available, fall back to openspec status --change "<name>" --json, tasks.md, and docs/superpowers/ file checks.

Resume rules:

  • On every context resume, rerun Step 0 and Step 1; do not trust conversation history for phase detection
  • If there is an active change and the worktree has uncommitted changes, handle them through comet/reference/dirty-worktree.md. That protocol defines checks, attribution, and prohibitions; this file does not repeat them
  • If phase: build, first check build_pause, plan, build_mode, and isolation (see details below):
    • If build_pause: plan-ready but isolation and build_mode are already set, treat as stale pause: first output [COMET] Detected stale pause (build_pause=plan-ready but isolation/build_mode already set), auto-clearing and continuing, then run "$COMET_BASH" "$COMET_STATE" set <name> build_pause null, then read the next unchecked task from tasks.md and resume execution per build_mode
    • If build_pause: plan-ready and the plan file exists, but isolation or build_mode is not yet set, return to the /comet-build plan-ready resume point, prompt the user to choose isolation and execution method, and do not regenerate the plan
    • If build_pause: plan-ready but the plan file is missing, return to /comet-build to handle corrupted state or regenerate the plan
    • If build_mode, isolation, or tdd_mode is unset, return to the corresponding /comet-build step to supplement before executing
    • If all are set, read the next unchecked task from tasks.md and continue:
      • If build_mode: subagent-driven-development, do not execute tasks directly in the main window; return to /comet-build's background subagent dispatch rules, main window only coordinates
      • Other execution modes follow /comet-build's corresponding rules
  • If phase: verify and verify_result: fail, enter the verification failure decision blocking point: pause and ask the user to fix or accept deviation; only after the user chooses fix, run "$COMET_BASH" "$COMET_STATE" transition <name> verify-fail and invoke /comet-build
  • If phase: open but proposal/design/tasks are complete, first run "$COMET_BASH" "$COMET_GUARD" <change-name> open --apply to repair state, then continue detection
  • If phase: archive, only invoke /comet-archive; /comet-archive must first wait for final archive confirmation. After archive succeeds, the change moves to the archive directory, so do not run guard against the old active directory

Step 2: Phase Determination (check in order, first match wins)

  1. archived: true or change moved to archive → Workflow complete
  2. verify_result: pass and archived is not true → Invoke /comet-archive (first perform final archive confirmation)
  3. verify_result: fail → Enter verification failure decision blocking point (pause and ask fix or accept deviation; only after user chooses fix, run verify-fail then /comet-build)
  4. phase: verify or tasks.md all checked → Invoke /comet-verify
  5. phase: build or has Design Doc but plan/execution incomplete → Route by workflow: hotfix → /comet-hotfix, tweak → /comet-tweak, full → /comet-build
  6. phase: design or has change but no Design Doc → Invoke /comet-design
  7. phase: open or active change exists but .comet.yaml is missing → Invoke /comet-open
  8. No active change → Invoke /comet-open

If metadata conflicts with file state, use verifiable file state as source of truth and correct .comet.yaml before continuing.

Preset Upgrade Criteria

hotfix → full (upgrade if any condition met):

  • Change involves 3+ files
  • Architecture changes (new modules, new interfaces, new dependencies)
  • Database schema changes
  • Fix introduces new public API
  • Fix scope exceeds a single function/module

tweak → full (upgrade if any condition met):

  • Change involves 5+ files
  • Cross-module coordination required
  • 5+ new test cases needed
  • Config item additions or deletions (not value changes)
  • New capability needed
  • Delta spec needed (existing spec affected)
Error Handling Quick Reference
ScenarioHandling
openspec list --json failsCheck if openspec is installed, prompt user to run openspec init
Sub-skill unavailableStop workflow, prompt to install or enable the corresponding skill
.comet.yaml malformed or missingUse file state as source of truth, correct with "$COMET_BASH" "$COMET_STATE" set then continue
Build/test failsReturn to build phase for fixes, do not enter verify
Incomplete change directory structureFill missing files according to comet-open artifact requirements
Phase Transitions
<IMPORTANT>
A single `/comet` invocation starts from the detected phase and advances to the next phase when exit conditions are met.

Flow chain: open → design → build → verify → archive

Continuous execution requirement: starting from the detected phase, the agent automatically continues through all later phases. But auto-advancing only applies at transition points without user decisions. When encountering user decision points, must use the current platform's available user input/confirmation mechanism to pause and wait for the user's explicit response. Must not use recommendation rules, defaults, or historical preferences to substitute for user confirmation, and must not just output a text prompt and then continue executing.

Distinguish phase advancement vs automatic handoff: each sub-skill runs phase guard --apply before exit to advance the .comet.yaml phase field. This step always happens and is not controlled by auto_transition. After that, the sub-skill runs "$COMET_BASH" "$COMET_STATE" next <name> to resolve the next action: when auto_transition is not false, output is NEXT: auto (auto-invoke next skill); when auto_transition is false, output is NEXT: manual (do not invoke next skill, show a manual run hint). Therefore auto_transition only controls next skill invocation, not phase advancement. Regardless of auto_transition, user decision points below remain blocking.

Decision points are blocking points: whenever reaching any of the following nodes, the current /comet invocation must stop, and follow the comet/reference/decision-point.md protocol to obtain the user's explicit choice. Only after the user explicitly chooses can the corresponding state fields be written and operations executed, then auto-advance resumes.

Nodes requiring user participation (pause only at these nodes):

  1. Open phase proposal/design/tasks review and confirmation
  2. Confirm design approach during brainstorming
  3. Plan-ready pause choice during build phase, followed by workflow configuration selection (isolation + execution method + TDD mode)
  4. Decide to fix or accept deviation when verify fails (including Spec drift handling)
  5. Choose branch handling method for finishing-branch
  6. Archive phase final confirmation before running the archive script
  7. Encounter upgrade conditions (hotfix/tweak → full workflow)
  8. Build phase scope expansion requiring redesign or new change split
  9. Open phase large PRD requiring confirmation to split into multiple changes

Agents should not skip these decision points; other unambiguous phase transitions must proceed automatically, must not exit midway. At decision points, must not skip user confirmation or choose automatically — must explicitly obtain the user's choice through the current platform's available user input/confirmation mechanism before continuing.

Red Flags — when these thoughts appear, STOP and check:

Agent ThoughtActual Risk
"The user would probably agree with this approach"Cannot decide for the user — use the current platform's user input/confirmation mechanism
"This is a small change, confirmation isn't needed"Decision points have no size exception — blocking points must wait
"The user chose A last time, so A again"Historical preference cannot substitute for current confirmation
"I explained the plan and the user didn't object"No objection ≠ consent — must use tool to get explicit choice
"The flow has reached this point, should be fine"Verification not passed ≠ passed — check verify_result
</IMPORTANT>

Show full SKILL.md (554 more words)Show less

Subcommand Quick Reference

CommandPhaseOwnerArtifacts
/comet-open1. OpenOpenSpecproposal.md, design.md, tasks.md
/comet-design2. Deep DesignSuperpowersDesign Doc, delta spec
/comet-build3. Plan and BuildSuperpowersImplementation plan, code commits
/comet-verify4. Verify and CloseBothVerification report, branch handling
/comet-archive5. ArchiveOpenSpecdelta→main spec sync, design doc markup, archive
/comet-hotfixPreset pathBothQuick fix (skip brainstorming)
/comet-tweakPreset pathBothSmall change (skip brainstorming and full plan)
/comet
  ↓ Auto-detect
/comet-open ──→ /comet-design ──→ /comet-build ──→ /comet-verify ──→ /comet-archive
  (OpenSpec)      (Superpowers)     (Superpowers)     (Both)          (OpenSpec)

/comet-hotfix (preset, skip brainstorming)
  open ──→ build ──→ verify ──→ archive
    ↑ If upgrade triggered → block for confirmation → supplement Design Doc → return to full workflow

/comet-tweak (preset, skip brainstorming and full plan)
  open ──→ lightweight build ──→ light verify ──→ archive
    ↑ If upgrade triggered → block for confirmation → supplement Design Doc → return to full workflow

Reference Appendix

State Machine Hard Constraints
  • Before build → verify, isolation must be branch or worktree
  • Before build → verify, build_mode must be selected
  • build_mode: subagent-driven-development must also have subagent_dispatch: confirmed
  • Before full workflow leaves build phase, tdd_mode must be selected as tdd or direct
  • build_mode: direct is allowed by default only for hotfix / tweak; full workflow requires direct_override: true
  • build_pause is not an execution method and must not be written to build_mode
  • These constraints are enforced by both comet-guard.sh build --apply and comet-state.sh transition <name> build-complete
.comet.yaml Field Reference

See comet/reference/comet-yaml-fields.md for complete field reference with examples and descriptions.

File Structure

See comet/reference/file-structure.md for the complete directory layout and artifact organization.

Auto-Transition Protocol

See comet/reference/auto-transition.md for the complete automatic handoff workflow.

Context Recovery

See comet/reference/context-recovery.md for structured recovery after context compression.

Decision Point Protocol

See comet/reference/decision-point.md for the complete user decision point protocol.

Debug Gate Protocol

See comet/reference/debug-gate.md for the complete debug gate protocol.

Script Location

Comet scripts are distributed in comet/scripts/. Do not hardcode paths — locate once, cache in env vars. This block is a standard boilerplate repeated in every sub-skill for independent loadability; changes must be kept in sync across all files (boilerplate version: v2, update this version when changing to help locate files needing sync):

bash
COMET_ENV="${COMET_ENV:-$(find . "$HOME"/.*/skills "$HOME/.config" "$HOME/.gemini" -path '*/comet/scripts/comet-env.sh' -type f -print -quit 2>/dev/null)}"
if [ -z "$COMET_ENV" ]; then
  echo "ERROR: comet-env.sh not found. Ensure the comet skill is installed." >&2
  return 1
fi
. "$COMET_ENV"

# Stop workflow when script location fails
if [ -z "$COMET_GUARD" ] || [ -z "$COMET_STATE" ] || [ -z "$COMET_HANDOFF" ] || [ -z "$COMET_ARCHIVE" ]; then
  echo "ERROR: Comet scripts not found. Ensure the comet skill is installed." >&2
  echo "Expected path pattern: */comet/scripts/comet-*.sh under project or platform skill directories" >&2
  return 1
fi

Auto state update: Guard supports --apply flag, automatically updating .comet.yaml state fields after checks pass:

bash
"$COMET_BASH" "$COMET_GUARD" <change-name> <phase> --apply

--apply delegates to comet-state transition. Use these semantic events when state changes need to be expressed directly:

bash
"$COMET_BASH" "$COMET_STATE" transition <change-name> open-complete
"$COMET_BASH" "$COMET_STATE" transition <change-name> design-complete
"$COMET_BASH" "$COMET_STATE" transition <change-name> build-complete
"$COMET_BASH" "$COMET_STATE" transition <change-name> verify-pass
"$COMET_BASH" "$COMET_STATE" transition <change-name> verify-fail
"$COMET_BASH" "$COMET_STATE" transition <archive-name> archived

Resolve next action: after guard-based phase advancement, use the next subcommand to determine whether to auto-invoke the next skill:

bash
"$COMET_BASH" "$COMET_STATE" next <change-name>

Output format: NEXT: auto|manual|done + SKILL: <skill-name> (omitted for done) + HINT (for manual only). With auto_transition: false, output is manual, which pauses only the next skill invocation and does not block phase updates.

Archive script: Complete all archive steps in one command:

bash
"$COMET_BASH" "$COMET_ARCHIVE" <change-name>

After loading comet, agents should run the variable assignments above once, then reuse $COMET_GUARD, $COMET_STATE, $COMET_HANDOFF, $COMET_ARCHIVE throughout the session.

Best Practices
  1. brainstorming cannot be skipped — Every change must undergo deep design (except hotfix and tweak)
  2. delta spec is a living document — Freely modify during phase 3, sync at archive
  3. Handoff packages are generated by scripts — OpenSpec → Superpowers context must be generated through comet-handoff.sh as compact traceable excerpts (use --full when needed), and validated by guard for source/hash/mode
  4. Keep tasks.md in sync — Check off each completed task
  5. Commit frequently — One commit per task, message reflects design intent
  6. Verify before archive confirmation — Enter /comet-archive only after /comet-verify passes, but wait for final user confirmation before running the archive script
  7. Classify incremental updates — Small edits, medium brainstorming, large new changes
  8. Plan must associate with change — File header contains change: and design-doc: metadata
  9. Archive closure — design doc and plan must mark archived-with status
  10. Modifying existing features — Just open a new change
  11. Preset has limits — Switch to full workflow promptly when hotfix/tweak meet upgrade conditions

© rpamis, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 16 other files (scripts) in eval/local/skills/benchmarks/039-release/comet-classic-039 of rpamis/comet.

  • SKILL.md
  • reference/auto-transition.md
  • reference/comet-yaml-fields.md
  • reference/context-recovery.md
  • reference/debug-gate.md
  • reference/decision-point.md
  • reference/dirty-worktree.md
  • reference/file-structure.md
  • reference/subagent-dispatch.md
  • rules/comet-phase-guard.md
  • scripts/comet-archive.sh
  • scripts/comet-env.sh
  • scripts/comet-guard.sh
  • scripts/comet-handoff.sh
  • scripts/comet-hook-guard.sh
  • scripts/comet-state.sh
  • scripts/comet-yaml-validate.sh

Open the folder on GitHubat commit 0fd42a0

Compare with similar skills

Comet next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Comet compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Comet this skillrpamis/comet3.2k—~4.4kAutomated safety check: PassMIT
MCP Server Builderanthropics/skills180k64 repos~2.3kAutomated safety check: PassApache-2.0
Hook Development for Claude Code Pluginsanthropics/claude-plugins-official38k11 repos~4.1kAutomated safety check: NotesApache-2.0
Using Superpowersfarm-fe/farm5.6k35 repos~1.4kAutomated safety check: PassMIT
Executing Plans Inlineobra/superpowers296k2 repos~5.1kAutomated safety check: PassMIT
Claude Code Agent Developmentanthropics/claude-plugins-official38k8 repos~2.8kAutomated safety check: PassApache-2.0

Similar skills

  • MCP Server Builder

    anthropics/skills

    Official

    Guides the design and implementation of Model Context Protocol servers in TypeScript or Python, from tool naming and error messages to evaluation.

    180k GitHub starsUsed in 64 repos~2.3k tokens
    Agent WorkflowsAuto-check passed
  • Hook Development for Claude Code Plugins

    anthropics/claude-plugins-official

    Official

    Explains how to write Claude Code plugin hooks, both prompt-based checks and bash commands, for events such as PreToolUse, Stop and SessionStart.

    38k GitHub starsUsed in 11 repos~4.1k tokens
    Agent WorkflowsAuto-check: notes
  • Using Superpowers

    farm-fe/farm

    A skill your agent uses when starting any conversation - establishes how to find and use skills, requiring Skill tool invocation before ANY response including clarifying questions

    5.6k GitHub starsUsed in 35 repos~1.4k tokens
    Agent WorkflowsAuto-check passed
  • Executing Plans Inline

    obra/superpowers

    Has the agent carry out an implementation plan itself, task by task in the current session, keeping a ledger, proving each step with a test and ending with one whole-branch review.

    296k GitHub starsUsed in 2 repos~5.1k tokens
    Agent WorkflowsAuto-check passed
  • Claude Code Agent Development

    anthropics/claude-plugins-official

    Official

    Explains how to write agents for Claude Code plugins: the markdown file with YAML frontmatter, trigger descriptions, model and color settings, and system prompt design.

    38k GitHub starsUsed in 8 repos~2.8k tokens
    Agent WorkflowsAuto-check passed
  • Skill Creator

    Azure/azqr

    Official

    Create new skills, modify and improve existing skills, and measure skill performance.

    795 GitHub starsUsed in 89 repos~8.2k tokens
    Agent WorkflowsAuto-check passed

More from rpamis/comet

All 52 skills in this repo
  • Comet

    rpamis/comet

    A skill your agent uses when 用户要启动或恢复 Comet 工作流,需要根据 active change、.comet.yaml、hotfix/tweak 意图路由到对应阶段 Skill。

    3.2k GitHub stars~2.7k tokensUpdated yesterday
    Auto-check passed
  • Comet

    rpamis/comet

    Comet workflow entry. An agent skill from rpamis/comet.

    3.2k GitHub stars~686 tokensUpdated yesterday
    Auto-check passed
  • Comet Archive

    rpamis/comet

    Archive and deliver a Classic change. An agent skill from rpamis/comet.

    3.2k GitHub stars~3.9k tokensUpdated yesterday
    Auto-check passed
  • Comet Classic

    rpamis/comet

    Comet Classic workflow entry. An agent skill from rpamis/comet.

    3.2k GitHub stars~3k tokensUpdated yesterday
    Auto-check passed
  • Comet Design

    rpamis/comet

    Complete the Classic technical design and obtain user confirmation.

    3.2k GitHub stars~4.3k tokensUpdated yesterday
    Auto-check passed
  • Comet Design

    rpamis/comet

    A skill your agent uses when full Comet change 已完成 open 阶段但缺少 Superpowers Design Doc,或 design 阶段需要从 OpenSpec 交接包恢复。

    3.2k GitHub stars~1.9k tokensUpdated yesterday
    Auto-check passed

Categories

Questions about Comet

What does Comet do?

Comet — OpenSpec + Superpowers dual-star development workflow. Comet is an agent skill from rpamis/comet. Comet — OpenSpec + Superpowers dual-star development workflow.

When should I use Comet?

Comet fits situations like: agent Workflows work in your project.

How do I install Comet in Claude Code?

Run `npx skills add rpamis/comet --skill comet -a claude-code`. Or copy the skill folder (eval/local/skills/benchmarks/039-release/comet-classic-039 in rpamis/comet) into .claude/skills/comet in your project. Claude Code loads it when a task matches its description.

How do I install Comet in Codex?

Run `npx skills add rpamis/comet --skill comet -a codex`. Or copy the skill folder (eval/local/skills/benchmarks/039-release/comet-classic-039 in rpamis/comet) into .agents/skills/comet in your project. Codex loads it when a task matches its description.

Can I use Comet in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add rpamis/comet --skill comet -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/comet, .gemini/skills/comet, .github/skills/comet and .opencode/skills/comet in your project.

What does Comet need to run?

Going by SKILL.md and its folder, Comet needs a shell for the scripts in its folder. Our summary lists: A Bash shell.

Does Comet access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Comet safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does Comet use?

Comet is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Comet use?

About 4.4k tokens (SKILL.md is roughly 17k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Comet?

Skills that share tags, products or a category with Comet: MCP Server Builder (anthropics/skills, 180k stars), Hook Development for Claude Code Plugins (anthropics/claude-plugins-official, 38k stars), Using Superpowers (farm-fe/farm, 5.6k stars) and Executing Plans Inline (obra/superpowers, 296k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Comet?

rpamis (a GitHub organization) maintains it in rpamis/comet, which has 3,160 GitHub stars. The repository holds 52 skills in this directory. The repository was last updated on October 8, 2026.

Source: rpamis/comet on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.