Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add settings/punt-labs/punt-kit/settingsgit clone --depth 1 https://github.com/punt-labs/punt-kitGrade C, and why
settings scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
"Bash(rm -rf /:*)", Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
"Bash(curl:*)", What it actually says
{
"permissions": {
"allow": [
"Bash(bash:*)",
"Bash(bd:*)",
"Bash(cat:*)",
"Bash(chmod +x:*)",
"Bash(claude mcp:*)",
"Bash(claude plugin:*)",
"Bash(export:*)",
"Bash(find:*)",
"Bash(gh:*)",
"Bash(git:*)",
"Bash(ls:*)",
"Bash(pip index:*)",
"Bash(punt:*)",
"Bash(python3:*)",
"Bash(sed:*)",
"Bash(shellcheck:*)",
"Bash(tail:*)",
"Bash(test:*)",
"Bash(uv:*)",
"WebFetch(domain:claude.com)",
"WebFetch(domain:docs.anthropic.com)",
"WebFetch(domain:elevenlabs.io)",
"WebFetch(domain:ics.uci.edu)",
"WebFetch(domain:survey.stackoverflow.co)",
"WebFetch(domain:venturebeat.com)",
"WebFetch(domain:www.anthropic.com)",
"WebFetch(domain:www.cartesia.ai)",
"mcp__github__*",
"mcp__plugin_biff_tty__*",
"mcp__plugin_github_github__*",
"mcp__plugin_quarry_quarry__*",
"mcp__quarry__*",
"Bash(uvx twine:*)",
"Read",
"Edit",
"Write"
],
"deny": [
"Bash(curl:*)",
"Bash(dd:*)",
"Bash(direnv allow:*)",
"Bash(ftp:*)",
"Bash(nc:*)",
"Bash(ncat:*)",
"Bash(netcat:*)",
"Bash(rm -rf /:*)",
"Bash(rm -rf ~:*)",
"Bash(scp:*)",
"Bash(socat:*)",
"Bash(ssh:*)",
"Bash(su:*)",
"Bash(sudo:*)",
"Bash(telnet:*)",
"Bash(tftp:*)",
"Bash(wget:*)",
"Edit(.env)",
"Edit(.envrc)"
]
}
}
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 64 lines scan C a805a2c6d7cf
settings is a settings file published in the GitHub repository punt-labs/punt-kit (1 stars, last pushed 2d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. A static security scan graded it C with 2 findings (recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other settings, from other repositories
prfaq
Agent settings declaring 113 allowed tools.
beadle
Agent settings declaring 1 hook event (PreToolUse) and 110 allowed tools.
quarry
Agent settings declaring 39 allowed tools.
z-spec
Agent settings declaring 114 allowed tools.
vox
Agent settings declaring 120 allowed tools.
biff
Agent settings declaring 118 allowed tools.