Coding agents superpack

Skill frankxai/awesome-hermes-agent-skills/skills/coding-agents-superpack

Awesome list of Hermes-compatible skills & plugins from across the web (+ small free open-core packs) | agents companion: frankxai/awesome-hermes-agents

Install

npx -y skills add frankxai/awesome-hermes-agent-skills --skill coding-agents-superpack

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use when orchestrating multiple coding agent CLIs (Claude Code, Codex, OpenCode, Grok CLI, Antigravity). Discovery, comparison, structured prompting, and multi-agent handoff patterns. Portable free pack — no host-specific secrets.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

4.0 KB, 864 tokens by cl100k_base, as published. Nobody here has run it

Coding Agents Superpack (Free)

Portable multi-CLI orchestration patterns for Hermes, Claude Code, Codex, OpenCode, Grok CLI, and similar agents.

This free pack is sanitized. Personal host quirks, private brand wiring, and machine-specific topology live in private overlays — not here.

When to Use

Multiple coding CLIs available; need discovery or comparison
Designing structured prompts for autonomous coding agents
Multi-agent collaboration with a shared board (Model Council pattern)
Building hybrid “orchestrator drives + specialist CLI executes” workflows

Don't use for: single-CLI deep flags (load that vendor’s skill/docs); private multi-brand empire ops.

Discovery (run first)

which codex && codex --version
which claude && claude --version && claude auth status
which opencode && opencode --version
which grok && grok --version
# health
codex doctor 2>/dev/null
claude doctor 2>/dev/null

Prefer project-scoped filesystem checks. Avoid recursive search from home or drive roots (slow, noisy, and unsafe on multi-device Windows setups).

Test agents in an isolated temp git repo:

TEST_DIR=$(mktemp -d)
cd "$TEST_DIR" && git init -q && echo "# test" > README.md && git add . && git commit -q -m init

Headless modes (automation)

Agent	Preferred non-interactive
Claude Code	`claude -p "…" --max-turns N --allowedTools "Read,Edit,Bash"`
Codex	`codex exec "…"` (sandbox flags per install)
OpenCode	`opencode run "…"`
Grok CLI	`grok --always-approve "…"` (confirm flags for your version)

Universal prompt structure

# Objective
Clear goal + measurable success criteria

# Context
1. Read these exact files first: …
2. Constraints: …

# Procedure
1. …
2. …

# Required Artifacts
- Files with exact paths
- Verification report

# Verification
Run tests/lint; report failures; do not claim done without evidence

Model Council (lightweight multi-agent)

When agents must collaborate (not isolated one-shots):

Shared board file: KANBAN_BOARD.md (or Hermes kanban)
Shared state: STATE.json or equivalent
Role prompts force: read board first, update board last
Explicit handoff artifacts (IMPLEMENTATION_BRIEF.md, etc.)

Roles example: Chair · Planner · Implementer · Reviewer · Researcher

Hybrid execution pattern

For large multi-repo work:

Orchestrator creates foundations with direct file/git ops (visible, fast)
CLI specialists get scoped tasks with --max-turns / budgets
Publish a single execution report that separates done vs delegated queue
Never leave silent open tasks without a handoff document

Safety & cost

Cap turns/budget on headless runs
Restrict tools to what the task needs
Prefer print/one-shot modes for CI
Clean up tmux/sessions after interactive work
Never put secrets in prompts or board files

Verification checklist

Discovery run; versions recorded
Prompt includes Objective/Context/Procedure/Verification
Headless flags set (max-turns/tools)
Artifacts exist on disk; tests/lint reported
Board updated if multi-agent

Related free resources

GenCreator / Starlight public OS repos on frankxai
Official Hermes skills: claude-code, codex, opencode, hermes-agent
Workflow Tier plugin (MIT multi-agent workflows)

License

MIT — Frank Riemer / GenCreator. Upstream CLI behavior belongs to their vendors.