Each watches a different moment of the same work; the recommended setup is all ten. Turn one off where it brings nothing.

Executive Self-Monitoring
executive-self-monitoring 1.11.2
Am I still on the plan?
Plan-anchored self-check: the agent re-reads the plan or ticket that defines the work, quotes its goal and its gate, names what pulls away from it, and keeps the plan current as decisions change it.
Read more

Epistemic Self-Monitoring
epistemic-self-monitoring 0.6.2
Is this actually verified?
Observation vs. conjecture: the agent keeps what it saw apart from what it concluded, and only a verified claim becomes a fact, a root cause or a closure.
Read more

Persistence Self-Monitoring
persistence-self-monitoring 0.5.3
Is it worth another try?
Persist-or-quit signal: counts repeated attempts on the same file, command or error, and asks the agent to switch approach or report to you when a threshold is crossed.
Read more

Termination Self-Monitoring
termination-self-monitoring 0.5.2
Is that a reason or a phrase?
Checkable reasons to stop: a stop on a feeling or a limit the agent does not manage - context, a long session, low confidence - is replaced by a checkable reason, or the work continues.
Read more

Coverage Self-Monitoring
coverage-self-monitoring 0.7.3
Did I do the hard part?
Parts-ledger discipline: the agent lists the parts of a request, takes the hardest first, and closes each one as done, blocked with a reason, or returned to you - none silently dropped.
Read more

Handoff Self-Monitoring
handoff-self-monitoring 0.8.2
Can the reader act on what I wrote?
Structured handoff: the agent's final message opens with a status, states the situation in your terms, writes any choice as options with a default, and ends with one next action.
Read more

Progress Self-Monitoring
progress-self-monitoring 0.10.2
Can the next session pick this up?
Cross-session ledger: what a session leaves blocked or returned, with the reason, and the next action, kept in .agent/progress.md and re-read before the next session works.
Read more

Integrity Self-Monitoring
integrity-self-monitoring 0.4.2
Is the result real?
Real-result discipline: when a service, a key or consistent tests are missing, the agent says so instead of shipping a fallback or made-up data that only looks done.
Read more

Aspiration Self-Monitoring
aspiration-self-monitoring 0.4.2
Did I review it?
Review before you close: the agent checks an edited result in the medium it will be used in - reads the document, runs the code, views the image - and compares it point by point with a source it did not write.
Read more

Hygiene Self-Monitoring
hygiene-self-monitoring 0.4.2
Did the change stay inside the request?
Change-scope self-check: the agent lists the public functions its change reaches, and puts back - or brings to you - any the request did not name.
Read more