Lightspeed

Speed-CLI agentic skill hub: index to install, protocol, commands, scripts, building blocks, function builder, orchestration, identity/SANS, MCP, security, o...

Security Scan

This is an instruction-only documentation hub for the Speed-CLI ecosystem; its declared requirements (none) align with its purpose, but following the linked docs may lead you to install and run third-party CLI/software that can access keys and run scripts — review before executing.

Purpose & Capability

The skill is a documentation/index hub for 'Speed-CLI' and related agentic skills. It declares no binaries, env vars, install steps, or credentials, which is coherent with an instruction-only README-style skill.

Instruction Scope

SKILL.md instructs agents to follow GitHub links, read per-skill docs, clone the speed-scripts repo, and run Speed-CLI commands such as `speed setup`, `speed start`, and `speed skill`. The doc itself does not instruct the agent to read or exfiltrate local secrets, but following the links may require running scripts or commands that use wallets, encrypted env, RPC endpoints, or private keys — those operations can have side effects and should be reviewed before execution.

Install Mechanism

No install specification is included in the skill package (instruction-only). The README mentions optional installs (e.g., an npm package) on the upstream project, which is a normal distribution method but would occur outside this skill and should be validated if chosen.

Credentials

The skill declares no required environment variables or credentials. However, the subject matter (MCP, keys, encrypted env, identity/SANS, wallets) implies that using the referenced tools will likely require wallet keys, RPC URLs, or other secrets. That is appropriate for the upstream tooling, but users should not provide such secrets to this skill itself without explicit need and verification.

Persistence & Privilege

The skill is not marked always:true and does not request persistent system-level privileges or modify other skills' configs. Autonomous invocation is allowed (platform default) but that does not appear to create an outsized privilege in this case.

Assessment

This skill is documentation-only and appears to be what it says. Before following links or running commands from the referenced repos: 1) Inspect the upstream GitHub repo and any scripts you plan to run; don't blindly execute clone + install + run sequences. 2) Never paste private keys/seed phrases into prompts; if you need to test, use a throwaway account or minimal-funds wallet. 3) If you install the CLI (e.g., via npm), verify the package name, version, and source and consider installing in an isolated environment. 4) Be cautious with any steps that mention encrypted env, key management, or MCP — those legitimately need secrets, but you should supply them only to trusted, audited tools. If you want, I can fetch and summarize the linked skill pages so you can review commands and required envs before running anything.

Speed-CLI — Agentic hub

Skill-style docs for the Speed-CLI multichain tool and speed-scripts: frontmatter, mental models, tables, pitfalls, and links. Full sources live in lightspeedfoundation/speed-skills on GitHub.

Browse on the web: GitHub Pages (repo root index.html).

Per-skill URL pattern:
https://github.com/lightspeedfoundation/speed-skills/blob/main/agentic/<category>/skill.md

---

How agents should use this folder

1. Match intent — use the description in each skill’s YAML frontmatter (trigger terms live there).
2. Read one skill end-to-end before mixing commands — Orchestration ties phases together.
3. Deep CLI flags — confirm with speed <cmd> --help; Commands summarizes roles, not every flag.
4. Scripts vs CLI vocabulary — Scripts points at scripts-reference; script-level flags and JSON are covered in Function builder (mirror of upstream function-builder-skill).

---

Skill index

Start

Install, speed setup, MCP, speed skill, whoami, scripting flags, advanced setup (Cursor + clone speed-scripts, e.g. cbBTC / trailing-stoploss-any).

Why Speed-CLI

Problems solved, design pillars, what the CLI is not, relation to speed-scripts.

Protocol

Env matrix, chains/explorers, SPEED address, 0x/Squid/oracles, RPC, MCP merge rules, identity/SANS.

Commands

Global flags, JSON contract, grouped subcommands, doctor behavior.

Scripts

Clone speed-scripts, folder discovery, *-skill.md, upstream doc links.

Building blocks

Base vs target, strategy families, regime thinking, hazards, decimals.

Function builder

Every speed flag/subcommand, JSON, decimals, PowerShell/bash patterns, pitfalls (mirrored from speed-scripts function-builder-skill).

Orchestration

Chaining JSON, multi-terminal, agent phases, errors, logging.

Identity & SANS

.speed on Base, profiles, SANS / OpenSea listings and offers.

MCP

speed start, mcp-url, encrypted env, trust model, debugging.

Security

Keys, allowances, revoke, MCP vs wallet boundary.

Observability

doctor, history, pending, xp.

---

GitHub Pages (maintainers)

Enable Settings → Pages → Branch: main, folder / (root) so the same repo serves the live site. Markdown under agentic/ is also viewable as raw .md on Pages if linked directly.

Optional later: per-command pages from --help, a fuller chain appendix, public-site risk disclaimers, pinned installs (npm install -g @lightspeed-cli/speed-cli@x.y.z) — none of these block publishing.