agentsclimarketplace

Agent context init

Skill lx-wnk/skills/skills/agent-context-init

Portable Agent Skills for AI coding agents (Claude Code, Codex, Cursor, Gemini) — agentskills.io-conformant, versioned, installable via skills.sh

Install
npx -y skills add lx-wnk/skills --skill agent-context-init

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

One thing to look at

  • 0 stars0 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

Initialize Agent-Context in the current project — sets up the layered context architecture with shared infrastructure, project-specific layers, memory stubs, and skill registry. Use this skill whenever the user asks to initialize Agent-Context, set up Agent-Context, bootstrap Agent-Context, or says things like "init agent-context", "set up context layers", "add agent-context to this project". Also use when the user wants to start using Agent-Context in a new or existing project.

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

3.5 KB, 635 tokens by cl100k_base, as published. Nobody here has run it

Initialize Agent-Context

Set up the Agent-Context layered architecture in the current project.

Examples

# Initialize with latest version
/agent-context-init

# Initialize with a specific version
/agent-context-init v0.2.1

Prerequisites Check

Before starting, verify this is a valid target:

ls .agent-context/.agent-context-version 2>/dev/null
  • If the file exists: STOP. Tell the user Agent-Context is already initialized and suggest using the update skill instead.
  • If the file does not exist: proceed with setup.

Setup

Resolve the target release tag first, then fetch the setup prompt pinned to that tag and follow its instructions. Fetching a specific release tag (never the mutable main branch) is deliberate: the tag is pinned for the whole run, so the fetched prompt is fixed between resolution and execution, and it changes only when a new release is published — unlike main, which every push mutates. (Full tamper-resistance additionally requires tag/release protection on the source repo.)

# Resolve target version: explicit $ARGUMENTS, else latest release (never mutable `main`)
TAG="${ARGUMENTS:-$(gh api repos/lx-wnk/Agent-Context/releases/latest --jq .tag_name)}"
curl -fsSL "https://raw.githubusercontent.com/lx-wnk/Agent-Context/${TAG}/.prompts/setup-prompt.md"

Follow the fetched instructions, subject to the consent gates in Trust Boundary & Consent below. The setup prompt handles:

  1. Version selection — fetch available releases, let user choose (or use $ARGUMENTS if a version was specified)
  2. Shared files — download tarball, extract framework files into .agent-context/
  3. Project templates — create project-owned files (AGENTS.md, layers, memory stubs, skills index)
  4. Agent sync — optionally install shared agents
  5. Plugin sync — merge plugins into .claude/settings.json
  6. Discovery scans — launch parallel subagents to detect tech stack and fill TODO placeholders

Trust Boundary & Consent

This skill fetches a remote setup prompt (pinned to a release tag, see above) and follows it. That prompt performs trust-expanding operations: it can install transitive agents, merge plugins, and modify .claude/settings.json. Treat these as privileged.

Before any step that installs agents (4), merges plugins (5), or writes to .claude/settings.json, stop and get explicit user confirmation — list exactly what will be installed or changed, then wait for a yes. Never install transitive agents or plugins silently. If the user declines a step, skip it and continue with the rest of setup.

After Setup

Report what was created:

  • Number of files created
  • Tech stack discovered
  • Any TODO placeholders that still need manual input
  • Remind user that project-owned files are theirs to customize

Gives 0 of the 12 instructions most context ai engineering skills give in 635 tokens

Counted across 1,193 of the 1,976 authors here whose files we hold, read 2026-08-06

  • dispatch a fresh implementer subagent per taskin 48 of 1193, across 19 files
  • dispatch final reviewer after all tasksin 37 of 1193, across 11 files
  • provide full task text to the subagentin 31 of 1193, across 10 files
  • review spec compliance before code qualityin 27 of 1193, across 10 files
  • make the hook script executablein 26 of 1193, across 8 files
  • re-snapshot after navigation or DOM changesin 25 of 1193, across 17 files
  • answer subagent questions before proceedingin 22 of 1193, across 7 files
  • mark task complete in TodoWrite after approvalin 22 of 1193, across 6 files
  • merge hook into existing settingsin 21 of 1193, across 3 files
  • read files before editing themin 21 of 1193, across 9 files
  • ask if installation is global or projectin 20 of 1193, across 2 files
  • copy the hook script to target locationin 20 of 1193, across 2 files

Said here and by no other author read

  • stop if agent-context already exists
  • fetch setup prompt for target release tag
  • never fetch from mutable main branch
  • get explicit user consent before merging plugins
  • get explicit user consent before modifying settings
  • never install agents or plugins silently

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.