feat(telemetry): missed-opportunity signal at checkpoint condensation

[jdx]

https://entire.io/gh/entireio/cli/trails/1071

Stacked on #2023.

Raw search-command counts can't say whether adoption is low — most sessions have nothing worth searching for. The denominator that makes the rate meaningful is: sessions that edited files carrying AI checkpoint history without ever consulting search. Entire is the only tool that can compute it, and until now nothing emitted it.

On each successful post-commit condensation, emit cli_checkpoint_condensed with two booleans and a count:

• used_search — substring probe over the already-in-memory transcript (canonical entire search and legacy entire checkpoint search spellings). Deliberately loose (a prompt merely mentioning the command matches); acceptable for an aggregate boolean, never used for behavior.
• prior_ai_history — did any committed file appear in a recent commit carrying an Entire-Checkpoint trailer? One bounded git log --skip=1 -n 50 --name-only subprocess; --skip=1 excludes the commit just made.
• files_committed — count only.

Content-free by construction: no file paths, prompts, or transcript content leave the machine. Gated on the opt-in telemetry setting before any work happens — the git-log probe never runs for opted-out users — plus ENTIRE_TELEMETRY_OPTOUT; the PostHog call is the existing detached-child path and never blocks the hook.

Tests: trailer-intersect semantics against a real temp repo (AI-touched vs human-touched vs HEAD-only files), transcript probe table test, payload shape + content-free assertions. Full suite: 9,170 unit / 492 integration / canary green.

🤖 Generated with Claude Code

Event name (added after review)

The event is cli_commit_condensed, not cli_checkpoint_condensed. A review
pass on the PR above this one read the absent emission on the doctor and
session-end condensation paths as data loss, which is the name's fault: it
promises "a checkpoint was condensed" and delivers "a commit condensed a
checkpoint".

Every property is commit-scoped — files_committed counts that commit's files,
and prior_ai_history asks whether commits before this one touched them,
which is what the probe's --skip=1 is for — so the event is named for the
commit, with the Go identifiers following. Renaming is free only while this PR
is unmerged: the event has never been emitted in a release and no dashboard
keys on it.

newCommitCondensedSignal now carries the scoping invariant: its sole caller is
condenseAndUpdateState, reached only from postCommitProcessSessionLocked.
The commit-less condensation paths deliberately emit nothing, because rows
without a commit are indistinguishable from genuine misses and would inflate
the denominator rather than complete it. Covering them would need a trigger
discriminator plus nullable commit-scoped fields — a change to what the metric
means, not a bug fix.

[Copilot]

Pull request overview

This PR adds a new, content-free telemetry signal emitted on successful post-commit checkpoint condensation to better measure “missed opportunity” adoption: sessions that committed files with prior AI checkpoint history without using entire search.

Changes:

• Introduces a cli_checkpoint_condensed telemetry event payload and detached sender (CheckpointCondensedSignal, builder, tracker).
• Adds strategy-side signal computation: used_search (substring probe over in-memory transcript) and prior_ai_history (bounded git log scan for recent commits with an Entire-Checkpoint trailer touching committed files).
• Wires emission into the post-commit condensation flow and adds unit tests for transcript probing and the git-log-based prior-history detector.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
cmd/entire/cli/telemetry/detached.go	Adds the new condensed-checkpoint telemetry signal type, payload builder, and detached tracking function.
cmd/entire/cli/telemetry/detached_test.go	Adds payload-building tests for the new event and centralizes the agent test constant.
cmd/entire/cli/strategy/telemetry_signals.go	Implements used_search detection + prior AI-history probing and emits the condensed-checkpoint telemetry signal (telemetry opt-in gated).
cmd/entire/cli/strategy/telemetry_signals_test.go	Adds unit tests for transcript search detection and prior-checkpoint-touch detection via a temp git repo.
cmd/entire/cli/strategy/manual_commit_types.go	Extends CondenseResult with UsedSearch for downstream telemetry emission.
cmd/entire/cli/strategy/manual_commit_hooks.go	Hooks telemetry emission into the successful condensation path.
cmd/entire/cli/strategy/manual_commit_condensation.go	Populates CondenseResult.UsedSearch from the extracted raw transcript.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[jdx]

Both findings addressed in 6f6eda2:

• Agent registry key: emitCheckpointCondensedTelemetry now resolves state.AgentType (display name, "Claude Code") to the registry key ("claude-code") via agent.GetByAgentType, so the agent property lines up with skill and command events. Unknown agent types fall back to the stored string rather than dropping the signal.
• Doc wording: CheckpointCondensedSignal now states the actual invariant — content-free metadata (no file paths, prompts, or transcript content) — and documents Agent as the registry key. Also applied the shared "auto" defaulting to this payload builder for consistency.

Rebased on the updated #2023.

[Copilot]

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated no new comments.

Suppressed comments (3)

cmd/entire/cli/strategy/telemetry_signals.go:93

• The environment opt-out is not checked until TrackCheckpointCondensedDetached, after this function has already run the git log probe. A user who opted in previously and then sets ENTIRE_TELEMETRY_OPTOUT still incurs the repository scan, contrary to the PR's guarantee that no work/probe occurs for opted-out users. Include the environment opt-out in this early gate.

	s, err := settings.Load(ctx)
	if err != nil || s.Telemetry == nil || !*s.Telemetry {
		return

cmd/entire/cli/strategy/manual_commit_hooks.go:1411

• This call executes inside the MutateSessionState callback at lines 1004-1010, so settings I/O, a git log subprocess, machine-ID lookup, and detached-process spawn all extend the per-session gate hold. The established telemetry pattern explicitly requires emission after mutation returns (cmd/entire/cli/lifecycle.go:1579-1581) to avoid blocking concurrent hooks. Capture the signal while locked, then run this telemetry work after MutateSessionState completes.

	// Content-free adoption signal (opt-in telemetry gated inside).
	emitCheckpointCondensedTelemetry(ctx, state, result)

cmd/entire/cli/strategy/telemetry_signals.go:40

• The acknowledged quoting behavior makes prior_ai_history incorrect for ordinary non-ASCII paths (for example, Git outputs "caf\303\251.go" while FilesTouched contains café.go), and line-based parsing also cannot represent filenames containing newlines. This systematically creates false negatives in the metric. Request NUL-delimited names (-z) and parse those names without trimming or Git quote decoding.

// Paths git quotes in --name-only output (e.g. non-ASCII names) won't match
// their unquoted FilesTouched form; that false-negative is acceptable for a
// telemetry boolean.

[jdx]

Addressed the three remaining review findings in 5a79a16, and rebased onto the updated #2023:

• Env opt-out checked before any probe work: emitCheckpointCondensedTelemetry now gates on the new telemetry.IsEnvOptedOut() before settings load and the git-log scan, so a user who opted in previously and then set ENTIRE_TELEMETRY_OPTOUT never pays for the repository probe. The other trackers share the helper.
• Emission moved outside the session gate: condensation now only snapshots a cheap, I/O-free condensedTelemetrySignal while locked; the PostCommit loop emits it after the session's MutateSessionState saves, alongside the skill-event emission — same pattern the previous round established on feat(telemetry): emit skill invocations as cli_skill_invoked events #2023.
• prior_ai_history no longer false-negatives on non-ASCII paths: the probe passes -z, so names arrive unquoted and NUL-terminated (café.go matches its FilesTouched form, and names containing newlines survive parsing). Pinned by a new test committing a non-ASCII path under a checkpoint trailer.

Full suite re-run on the stack: 9,324 unit / 514 integration / canary green.

[jdx]

@evisdren sorry — I pushed after your approval and dismissed it. Here's what changed so you can re-review just the delta.

One new commit: 26c4cd871, a rename. cli_checkpoint_condensed → cli_commit_condensed, with the Go identifiers following (CommitCondensedSignal, BuildCommitCondensedPayload, TrackCommitCondensedDetached, emitCommitCondensedTelemetry, newCommitCondensedSignal, commitCondensedSignal).

Why. A review pass on #2100 filed two HIGH findings reading the absent emission on the doctor and session-end condensation paths as data loss. That reading is the name's fault: the event promises "a checkpoint was condensed" and delivers "a commit condensed a checkpoint". Every property is commit-scoped — files_committed counts that commit's files, and prior_ai_history asks whether commits before this one touched them, which is exactly what the probe's --skip=1 is for. So the event is now named for the commit that is part of its identity.

Why now. Renaming is free only while this PR is unmerged: the event has never been emitted in a release and no dashboard keys on it. After merge it stops being free, which is why I didn't leave it as a follow-up.

newCommitCondensedSignal also now carries the scoping invariant — its sole caller is condenseAndUpdateState, reached only from postCommitProcessSessionLocked. The commit-less condensation paths deliberately emit nothing: rows without a commit are indistinguishable from genuine misses and would inflate the denominator rather than complete it. Covering them would need a trigger discriminator plus nullable commit-scoped fields — a change to what the metric means, not a bug fix. Both findings are dismissed with that rationale recorded.

Review shortcut: git diff 279f54263 26c4cd871 is the whole delta since your approval — 4 files, and apart from comments the only non-mechanical line is the event-name string itself.

Lint clean, test:ci green (unit + integration + canary).

[Soph]

used_search is a raw substring probe, and on real transcripts it is dominated by false positives.

Measured over 328 local Claude Code transcripts older than 3 days: 18 would set used_search=true; 1 actually ran the command (Bash tool_use command match).

The false positives are structural:

• setup_search_skill.go:96,105 puts entire search --json in the search skill's description and body — Entire installs the artifact that trips its own probe (3 of the 18).
• investigate/prompt.go:84 injects entire search "<phrase>" --json into every investigate prompt.
• Any session that reads this repo's own source (search_cmd.go, checkpoint_group.go, explain.go, cell_target.go) or merely discusses the command.

The body calls the looseness acceptable for an aggregate boolean, and I agree with that in principle — but the direction is the problem. Inflating used_search deflates the missed-opportunity rate, so the metric reads "adoption is fine" for the wrong reason, with nothing in the data to show it.

Same-cost fix: match an actual tool invocation rather than a substring. The transcript is already in memory and already walked for tool_use blocks (claudecode.ExtractSkillEvents); matching a Bash command field against entire (checkpoint )?search is what produced the "1" above.

Three smaller findings (per-session git log, the reintroduced telemetry-gate idiom, %B delimiter escaping) are on trail #1071.