> ## Documentation Index
> Fetch the complete documentation index at: https://docs.stet.sh/llms.txt
> Use this file to discover all available pages before exploring further.
# Change control for AI coding agents
> Change control for AI coding agents
Stet replays accepted repository work so you can measure whether a change to an
`AGENTS.md` file, shared skill, model, reasoning setting, or tool policy is safe
to keep. Your coding agent proposes and applies changes; Stet evaluates matched
work and returns a scoped decision.
## Find your next answer
Compare current and proposed instructions on the same retained tasks.
Iterate one instruction lever at a time with a bounded stop rule.
Test a skill against an appropriate absent or committed baseline.
Compare configurations on the same repository task slice.
## The loop
Accepted repository work becomes replayable tasks. Stet runs a baseline and a
candidate against the same task slice, evaluates their patches with tests and
graders, and writes a Trial Result with one scoped recommendation.
The evaluator is separate from the coding agent that made the change. That
separation keeps the question clear: did this bounded harness change improve
behavior on the declared work, and is the evidence strong enough for the next
action?
## Example Trial Result
This is a historical April 2026 model comparison across 28 paired Zod tasks,
with Opus 4.6 and Opus 4.7 both using high reasoning. It is a scoped example,
not a current or universal model ranking.
| Observed measure | Opus 4.6 | Opus 4.7 |
| ---------------------- | -------: | -------: |
| Test pass rate | 42.9% | 42.9% |
| Equivalence rate | 32.1% | 46.4% |
| Observed cost per task | \$19.96 | \$8.11 |
| Mean agent duration | 7m 58s | 3m 12s |
**PROMOTE** candidate **Opus 4.7**, high confidence, for this recorded
harness and task corpus.
On this task corpus, the historical receipt held observed test pass rate steady,
showed higher observed equivalence, and used less observed cost and time. The
receipt also used declared grader evidence that is not reproduced here.
The scope and evidence limits are part of the result:
* Historical April 2026 result, scoped to this 28-task Zod corpus and recorded harness.
* The legacy report predates Stet's current calibration and claim-readiness fields.
* The displayed metric values are observations, not generalized or uncertainty-calibrated improvement claims.
* The receipt also used declared grader evidence not reproduced here.
```json theme={null}
{
"schema_version": "stet.receipt/v1",
"source_commit": "e1980547",
"report_sha256": "88ca838e3098ae99043bb7d71454e608490938e05f17116352a929a4c2f8bf74",
"historical": true,
"historical_period": "April 2026",
"corpus": {
"name": "Zod",
"task_count": 28,
"paired": true,
"invalid_task_count": 0
},
"baseline": {
"model": "Opus 4.6",
"reasoning": "high"
},
"candidate": {
"model": "Opus 4.7",
"reasoning": "high"
},
"decision": {
"recommendation": "PROMOTE",
"candidate": "Opus 4.7",
"confidence": "high"
},
"metrics": [
{
"name": "Test pass rate",
"baseline": "42.9%",
"candidate": "42.9%"
},
{
"name": "Equivalence rate",
"baseline": "32.1%",
"candidate": "46.4%"
},
{
"name": "Observed cost per task",
"baseline": "$19.96",
"candidate": "$8.11"
},
{
"name": "Mean agent duration",
"baseline": "7m 58s",
"candidate": "3m 12s"
}
],
"caveats": [
"Historical April 2026 result, scoped to this 28-task Zod corpus and recorded harness.",
"The legacy report predates Stet's current calibration and claim-readiness fields.",
"The displayed metric values are observations, not generalized or uncertainty-calibrated improvement claims.",
"The receipt also used declared grader evidence not reproduced here."
],
"publication_approved_at": "2026-07-14",
"publication_approved_by": "benredmond"
}
```
Read [how Stet works](/concepts/how-stet-works) for the mechanism, then use
[Read a Trial Result](/concepts/trial-result) when you have a result to
interpret.
## Start with a bounded first session
[Follow the quickstart](/quickstart) to install the CLI and skill, verify your
machine, onboard one repository, and stop with an onboarding receipt before an
expensive evaluation.