Telemetry and auditing

Measuring where context goes, where agents get blocked, and whether gates are earning their friction.

Everything here is local. It reads your own session transcripts on your own machine and writes to your own filesystem. Nothing is transmitted.

Context composition

df-tools context --limit 150
df-tools context --root /path/to/projects --raw

Recomputes where the context window actually goes, from your own transcripts: tool results versus tool-call inputs versus assistant text versus images, plus per-tool call counts and averages.

Images are priced per block

df-tools context prices images at roughly 1,500 tokens per block rather than by base64 length. Counting base64 characters overstates image cost by about 25× — that was the one real error in the original context audit.

See context discipline for what to do with the numbers.

Session audit

df-tools session-audit --limit 150
df-tools session-audit --since 2026-08-01

Classifies blocking events across sessions: where agents got stuck, what stopped them, and how often. This is the acceptance test for the gate-correctness and model-tier work — “this gate false-positives” is either in this output or it is an anecdote.

Telemetry view

df-tools telemetry
df-tools telemetry --scan --limit 150

One status-facing view with advisories, combining planning state with session analysis. --scan includes a fresh session audit rather than cached data.

Transcript export

df-tools transcript-export
df-tools transcript-export --out ~/devflow-index.jsonl --full ~/transcript-archive --limit 500

Writes a compact per-session index before retention deletes the underlying transcripts. Default output is ~/.claude/devflow/transcript-index.jsonl.

Run it periodically if you care about longitudinal measurement — transcripts age out, and the index is what survives.

Override log

df-tools override --gate gate-edits --reason "hand-fixing a generated file"
df-tools override --list --limit 20

Structured, logged gate overrides. The point is not bureaucracy — a gate that is overridden constantly is a gate that is wrong, and this log is the evidence that makes the case instead of an argument about it.

The audit log

The verify-completion Stop hook emits a JSONL entry per completion to ~/.claude/devflow/audit.log, or to DEVFLOW_AUDIT_LOG_PATH if set. It is best-effort and never blocks the Stop event.

Benchmarks and duplication

df-tools benchmark          # timing across recorded runs
df-tools dup-detect         # duplicated planning artifacts
df-tools survey decimal-objectives --root <path>