Critical-risk tools in Nodebench
9 of the 724 tools in Nodebench are classified as critical risk. This page profiles those tools specifically, with recommended policy actions and the attack patterns that target them.
Every operation listed below is an action PolicyLayer recommends controlling at the transport layer. Open any tool to see the full profile, risk score, and YAML policy snippet.
Tools at critical risk
-
compress_or_expand_textDestructivePrecisely compress or expand academic text by a target word count. Compress mode: remove filler words, convert clauses to phrases, passive to active. Expand mode: add logical co...
-
delete_learningDestructiveDelete a learning by key. Use when a learning is outdated or incorrect.
-
dive_changelogDestructiveRecord a change made to fix a bug, design issue, or improve a component. Links before/after screenshots to show what changed visually. Optionally references git commits and chan...
-
founder_local_weekly_resetDestructiveOne-call convenience: gathers all local context and produces a complete
-
founder_weekly_resetDestructiveFull weekly reset: discover → load → synthesize → track → milestone
-
remove_ai_signaturesDestructiveDetect and remove AI-generated writing signatures from academic text. First runs pattern matching for known AI vocabulary (leverage, delve, tapestry, mechanical connectors), the...
-
share_revoke_packet_linkDestructiveRevoke a local share link so it no longer counts as active.
-
trackDestructiveAdd, check, remove, or list tracked entities with one workflow tool. The default path optimizes for watched entities that can feed future nudges and brief artifacts.
-
unload_toolsetDestructiveRemove a dynamically loaded toolset from the current session to free up context. Cannot unload toolsets from the initial preset.
Attacks that target this class
Critical-risk tools in any server share these documented attack patterns. Each links to the full case and the defensive policy.