Last updated: 2026-07-19
trigger-tree sends no data off your machine. Its local telemetry stays in your project directory, under your control. There are no network calls, external services, analytics vendors, or runtime dependencies.
trigger-tree’s hooks write a local event log to $PROJECT/.trigger-tree/history.jsonl
inside the project you use it in:
rg/grep/find commands, or
consumed by successful Bash cat/head/tail/non-mutating sed/awk calls.
Bash reader arguments are checked after shell variables, substitutions, loops, and
globs resolve; only matching normalized paths are retained, with event metadata.TT_LOG_PROMPTS in .trigger-tree/config.sh:
truncate (setup default) — first 200 characters, local and gitignoredhash — a SHA-1 digest only, no text; also the fallback before setupoff — a bare marker, nothing else/tt note.The optional HTML report (/tt insights) is generated locally to
.trigger-tree/report.html. If you choose to publish it as a Claude Artifact, that
is an explicit action you take through Claude Code — the plugin itself never uploads
anything.
You own the data. Delete .trigger-tree/ in your project at any time to remove all
recorded history. The directory is gitignored by the recommended setup, so it is not
committed or shared through version control.
Changes to this policy are recorded in this file’s git history. Questions: me@heddevanderheide.nl.