Sensors

Watches the agent's own tool calls and sessions, and reports the patterns that mean it is in trouble.

type computational tools 2 loops 1

Compares the public interface of a change with the last release, and reports every change that breaks a caller.

type computational tools 2 loops 1

Drives a scripted flow through a page in a real browser, and reports the step that broke.

type computational tools 5 loops 1

Compiles the code or checks its types, and reports each error with its place.

type computational tools 3 loops 1

Reads the results of the project's CI for a commit: which jobs passed, which failed, and the failing step's log.

type computational tools 2 loops 2

Shows what the tests do not check: lines no test runs, and changes to the code that no test notices.

type computational tools 2 loops 0

Measures how work flows through the project over weeks: how long changes wait, how soon they get a first review, how old the backlog is, how much CI costs.

type computational tools 1 loops 1

Measures the graph of imports between modules (cycles, fan-in, fan-out) and compares it with an approved baseline.

type computational tools 1 loops 1

Reads the errors real users hit, grouped into issues with stack traces and counts, and whether they came back after a fix.

type computational tools 1 loops 2

Runs an agent, a prompt or a skill many times on cases with known answers, and reports how often it gets them right.

type computational tools 2 loops 1

Feeds the code generated inputs or broken dependencies, and reports the smallest case that makes it fail.

type computational tools 5 loops 1

Checks from inside an agent's environment that it is isolated: no network it should not reach, no secrets it should not see, nothing left after teardown.

type computational tools 1 loops 1

Checks code against rules, and names each violation with its place and how to fix it.

type computational tools 5 loops 2

Queries the logs of a run or a release, and reports the errors in them, grouped and counted.

type computational tools 1 loops 4

Reads the record of what the loops caught and missed (escapes, corrections, releases not confirmed) to find the loop that should have caught a problem earlier.

type computational tools 2 loops 1

Checks one hypothesis on real values: a call in a REPL, a breakpoint in a debugger, a request with an HTTP client.

type computational tools 4 loops 1

Queries the service's metrics (error rates, latency, resource use) and compares them with a limit, an SLO or the previous version.

type computational tools 2 loops 1

Measures time, memory, size or load against a baseline or a budget, and reports what got worse.

type computational tools 7 loops 1

Reads the repository's state and diff directly, instead of taking the agent's word for what changed.

type computational tools 1 loops 0

Captures what the interface looks like, as an image or a video, for a comparison or as evidence.

type computational tools 3 loops 1

Scans code, history, dependencies or a running app for known security problems and licence breaches.

type computational tools 7 loops 1

Starts the built artifact or reaches the deployed release, and checks that it answers and that its main paths work.

type computational tools 6 loops 3

Checks a file that has a required shape (a config, a schema, a CI workflow, a spec, a changelog, a release artifact) against that shape or a policy.

type computational tools 5 loops 1

Drives a terminal app through a scripted session at a fixed size, sending keys and reading the rendered screen.

type computational tools 1 loops 0

Runs the project's tests on the changed code and reports which pass and which fail.

type computational tools 5 loops 3

Queries the traces of requests through the system, and reports the spans that failed or took too long.

type computational tools 1 loops 2

A model scores a result, a screenshot, a sample of logs or the risk of an action against a rubric.

type inferential tools 3 loops 2

A separate agent with a fresh context reads a diff, a document or some code against a checklist or a policy, and lists what it finds.

type inferential tools 1 loops 3

The agent asks a person a structured question (approve a plan, choose an option, confirm a risky action) and gets an answer it can act on.

type human tools 4 loops 2

A person looks at the result (the diff, a demo, a sample of the agent's work, the product itself) and says whether it is what was wanted.

type human tools 2 loops 2

A person follows the agent's work as it happens, and interrupts when it drifts.

type human tools 1 loops 0

Reads what real users say and do (feedback, support tickets, experiment results) in aggregate, with counts.

type human tools 2 loops 0