Agent hook configuration access
A call refers to Claude Code settings and includes hook-related text.
Why it is risky
Section titled “Why it is risky”Hooks can run commands automatically around agent activity. An unauthorized change can install recurring execution or weaken monitoring.
Example
Section titled “Example”An agent adds a hook that sends environment data to a service each time a tool runs.
This is a fictional example, not an event from a customer.
What Omnodex Detects
Section titled “What Omnodex Detects”The rule combines a matching settings path with the word hooks in captured parameters. The current path condition does not require a write.
| Rule ID | Event category | Tier | Base severity |
|---|---|---|---|
RULE_SUPPLY_CHAIN_HOOK_CONFIG_WRITE | supply_chain | Community | HIGH |
Base severity is the rule default. A finding involving content saved for later execution may have a lower severity. Your loaded rule version and event details determine the finding you see.
What to check next
Section titled “What to check next”Inspect the before-and-after settings and each hook command. Verify whether a write occurred and who authorized it.
False positives
Section titled “False positives”Reading settings for an authorized hook review can match. Installing a legitimate formatter or monitoring hook can also trigger it.
False negatives
Section titled “False negatives”Other platforms or hook paths, unsupported parameter fields, and subprocess edits that are not represented in the tool call may be missed.
Technical details
Section titled “Technical details”Despite the rule ID ending in WRITE, the finding can arise from a read or path reference when the hook keyword is present.
A finding describes a pattern in captured activity. Review the original call, its result, and the task context before concluding that harm occurred. Absence of a finding does not establish that an operation is safe.