compliance-hooks-setup — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited compliance-hooks-setup (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
This plugin populates the empty hooks/hooks.json scaffold that Anthropic's vertical finance plugins (private-equity, financial-analysis, investment-banking, equity-research, wealth-management) ship with.
# mnpi:cleared marker.mcp__ololand__generate_investment_memo, generate_cim, export_deal_dossier, plus all Write/Edit/MultiEdit). Scans for $-amounts, %, and multiples without an adjacent citation. Warns by default; set OLOLAND_CITATION_BLOCK=1 to upgrade to a hard deny.~/.ololand/provenance/YYYY-MM-DD.ndjson after every generative tool call. The ledger stays local; server-side MCP rail auditing is authoritative for OloLand tool calls.mcp__ololand__*). Mirrors every OloLand MCP call into ~/.ololand/audit/YYYY-MM-DD.ndjson with phase, tool name, and a 2KB payload head.Anthropic's private-equity/hooks/hooks.json ships as []. Same for the other four vertical finance plugins. There is no compliance hook, no citation enforcement, no provenance writeback, no MNPI guard. For a regulated buy-side workflow this is unacceptable — you cannot defend a number to the IC if you cannot prove what ran, what it was sourced from, or that no MNPI leaked into a prompt.
| Var | Default | Purpose |
|---|---|---|
OLOLAND_CITATION_BLOCK | 0 | Set to 1 to make the citation enforcer block (exit 2) rather than warn. |
OLOLAND_AGENT_KEY | unset | MCP connector authentication for headless clients; these hooks do not send it anywhere. |
Both ledgers live in ~/.ololand/:
~/.ololand/audit/YYYY-MM-DD.ndjson # every mcp__ololand__* call
~/.ololand/provenance/YYYY-MM-DD.ndjson # every generative outputInspect with jq -c '.' ~/.ololand/provenance/$(date -u +%Y-%m-%d).ndjson. The OloLand MCP rail separately records server-side tool-call audit events.
Cursor and Grok Bot do not execute Claude PreToolUse / PostToolUse matchers. Do not claim the MNPI guard, citation enforcer, or local NDJSON ledgers are armed on those hosts. A SessionStart banner may still run from the Cursor-format hooks file. For Cursor / Grok Bot, treat this skill as an explicit, user-invoked review:
If authorization is unclear, stop and request an authorized human review. A # mnpi:cleared marker does not establish permission.
documents. Preserve document/page references and label unverified claims. Never invent citations or turn missing evidence into a pass.
server actually handled. This review does not write ~/.ololand/audit or ~/.ololand/provenance ledgers unless the user asks you to append them.
This plugin is designed to run alongside ololand-dd, Anthropic's vertical finance plugins, and any custom plugin a firm installs. On Claude, hooks compose additively — multiple plugins' PreToolUse matchers all run for the same event. On Cursor / Grok Bot, compose by installing the plugins and running this explicit review skill.
claude plugin list shows ololand-compliance-hooks enabled. Run bash ~/.claude/plugins/cache/ololand-compliance-hooks/*/scripts/mnpi_guard.sh < /dev/null to test the script directly.~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.