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 hooks/zircote/human-voice/session-startgit clone --depth 1 https://github.com/zircote/human-voiceGrade A, and why
SessionStart scanned grade A with 0 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 commands this hook runs, not an audit. A hook is shell that executes on your machine at the event it names, which is why every command in it is printed with what was found.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
{
"SessionStart": [
{
"hooks": [
{
"type": "command",
"command": "DATA_DIR=\"$HOME/.human-voice\"; mkdir -p \"$DATA_DIR\"; [ -f \"$DATA_DIR/observer-protocol.md\" ] || cp \"${CLAUDE_PLUGIN_ROOT}/templates/observer-protocol.md\" \"$DATA_DIR/observer-protocol.md\" 2>/dev/null; cd \"${CLAUDE_PLUGIN_ROOT}\" && python3 -c \"from lib.profile_registry import resolve_active_profile; resolve_active_profile('$(pwd)')\" 2>/dev/null; true"
},
{
"type": "command",
"command": "echo \"Read $HOME/.human-voice/voice-prompt.txt and $HOME/.human-voice/observer-protocol.md if they exist. The first defines the user'\"'\"'s voice profile rules for content generation. The second defines a passive observation protocol: silently note patterns in the user'\"'\"'s own writing (not pasted text, not AI output) and compare against the profile. Follow both silently. Do not announce that you loaded them.\""
}
]
}
]
}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 · 16 lines scan A d181fcf6f3bc
SessionStart is a hook published in the GitHub repository zircote/human-voice (10 stars, last pushed 1mo ago), licensed MIT. Its token cost is not measured: a hook is shell that never enters the context. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other hooks, from other repositories
PostToolUse
Runs after a tool call finishes for Edit, Write, MultiEdit, Bash, Edit, Write, MultiEdit, Edit, Write, MultiEdit, Write, Edit and MultiEdit tool calls, executing token-guard.py, claude-path-write-guard.py, claude-integrity-tripwire.py, prompt-only-enforcement-guard.py, enforced-claim-lint.py, wiring-check.py…
PreCompact
Runs before the context is compacted, running an inline shell check. From luanmorenommaciel/brief-spec.
PostToolUse
Runs after a tool call finishes for Edit and Write tool calls, executing rust-quality.sh, file-size-guard.sh, catalog-sync-guard.sh, doc-sync-guard.sh, markdown-size-guard.sh, semantic-linebreak-guard.sh and section-ref-guard.sh (7 commands). From outlmd/outl.
SessionStart
Runs when a session starts on compact and resume, executing exomem-continuation-checkpoint.sh via bash. From Artexis10/exomem.
UserPromptSubmit
Runs when you submit a prompt, before the agent sees it, executing context-tracking.sh. From agenticnotetaking/eidos.
SessionStart
Runs when a session starts, executing index-sessions.sh via bash with --recent. From bashiraziz/llm-wiki-template.