Kira runs on your own machine — the client starts it, so there is no endpoint to ping. 58 installs a week from npm. Last commit 5 Aug 2026.
Skills and Scars for AI agents - your agent records its failures and never repeats them.
Every tool that appeared, vanished or quietly changed what it asks for. Recorded since 9 September 2026. No other catalogue keeps this.
Today is the operative word: we check Kira every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add kira -- npx -y kira-mcp
{
"mcpServers": {
"kira": {
"args": [
"-y",
"kira-mcp"
],
"command": "npx"
}
}
}
[mcp_servers.kira]
command = "npx"
args = ["-y", "kira-mcp"]
{
"mcpServers": {
"kira": {
"args": [
"-y",
"kira-mcp"
],
"command": "npx"
}
}
}
{
"mcpServers": {
"kira": {
"args": [
"-y",
"kira-mcp"
],
"command": "npx"
}
}
}
This one needs environment variables set before it will start:
KIRA_TELEMETRY (Override telemetry consent level for the process: off | basic | full. Default: basic on first run, with detail (note/context) gated by explicit opt-in. Full wire format and redaction rules in PRIVACY.md.), KIRA_TELEMETRY_URL (Override the telemetry ingest endpoint. Default: https://kira-telemetry.workers.dev/v1/reports.), KIRA_HOME (Where consent state and the local report log live. Default: ~/.kira), KIRA_PRO_KEY (Optional ES256-signed JWT for the Kira Pro tier (real-time skill/scar updates from CDN). Free tier works without this.).
The author declared them in the registry entry; get the values from the project itself.
Give your AI agent an identity: its own email, webhook, repo, secrets vault, memory and skills.
The static analyzer your AI agent reads — fix-ready, machine-readable scan reports over MCP.
Generate and validate a .specs/ bundle for your repo, then hand it to your AI coding agent
Abnormal Security email threats, cases, and reporting in your terminal and your AI agents.
Design your site's AI avatar in the ASMI studio; your coding agent ships it into your repo.
Toolbox for AI agent skills: scaffold, lint, pack, install, and serve SKILL.md skills
A package manager for Agent Skills — discover, vet, and pin them across Claude Code and Codex.
Persistent memory for AI coding agents — past bugs, decisions, and your corrections, in your repo.
Answers built from our own checks of this server.