Agent skill · laurigates
tool-result-traps
Tool results that mean something other than they look — a pattern that never compiled, a glob that cannot match a directory, one tier searched but reported absent, a rejected flag reading as "no results". Use when an empty result is about to gate an action, land in a doc, or be reported as done.
What it needs
About 12k tokens when loaded.
What this skill does
Tool Result Traps Promoted from the always-loaded ~/.claude/rules/tool-use-patterns.md rule, whose stub points here. The section bodies below are that rule's verbatim text. One law across all of them: an empty result, a green exit, and a well-formed line of output are each claims about mechanics, not about content. Every trap here produces output that is indistinguishable from a correct answer — no error, no warning, nothing on stderr — and the damage lands when that output is used as a verdict: "no duplicate exists", "the rename is complete", "nothing was lost", "the agent got its input". In one case the diagnostic was written — and the caller's own 2>/dev/null is what made it disappear. The recurring fix is equally uniform: control-test any negative that gates an action. Re-run the same command shape against something you know is present. If the control also comes back empty, the tool is broken, not the tree clean. One control run is cheap; a wrong negative is a confidently-reported non-finding. When to Use This Skill Use this skill when... Skip when... --- --- A zero-match or empty result is about to be reported as "clean", "complete", or "none found" The result is non-empty and you are acting on what it contains Deduping before filing an issue, or concluding nobody reported something Reading a single known file whose path you just listed Declaring a bulk rename, migration, or sweep finished Mid-sweep, still transforming matches A verification loop's input set is built from a relative path The path was resolved absolutely in the same command that used it Batching parallel tool calls whose siblings may exit non-zero A single call, or a batch of confirmed-present paths A Workflow script's agents report thin or generic findings on rich material Agents are returning specific, grounded detail An rg/git grep result contradicts something you read directly Output matches an independent read of the same file Concluding a skill, command, recipe, or binary "does not exist" …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills laurigates/tool-result-traps