Mixed freshness. Depending on the panel, rows here are read live from elisa-core or come from a frozen copy of ELISA’s local store taken on 22 October 2026 — each panel says which, underneath itself. No agent is running.

ELISA
Connecting

Agent 2

Checks that a job claiming to be done has the evidence it asserts, and that each external claim in its output is actually true.

Agent 2

Not connectedIdle — nothing queued

Evidence & claim verification · Oversight (not in the ELISA spec)

oversight.agent_2

Profile charter name: Pramana — a label from the Hermes profile. Nothing routes on it; routing is by id only.

Measured — state right now

Idle — nothing queued. A profile exists, its queue is empty and nothing is pending. Idle describes the queue, not a live process. This UI is not connected to a running agent and never starts one.

Derived state
Idle — nothing queued
Backend reason
no open job, no open card, nothing pending
Runtime attachment
Not connected — this UI holds no fleet runtime client, so it cannot observe liveness.
Hermes profile
oversight-a2
Profile built
Yes
Runtime lock
Not locked · sector dispatchable: yes
May create its own cards
No
Spec source
specs/governance.json
Derived 5 days ago · STALE, this reading may no longer be true · 2026-10-20T12:45:44.000Zfrom elisa.db:agents.runtime_locked+sectors.runnable (via dispatchable()) · elisa.db:approvals.status='pending' · elisa.db:jobs.status · kanban.db:tasks.status (assignee=profile)

Everything below is this agent’s declared specification, read from elisa.db. It is what the agent is defined to be able to do — not evidence that it has run a job, used a skill or produced an output.

Mission (declared)

Checks that a job claiming to be done has the evidence it asserts, and that each external claim in its output is actually true.

Will not (declared, and enforced at the gate)

  • write, deploy, promote, roll back or re-run anything on any connected service; every call it makes is read-only (script-checkable: block every non-read tool name on every server for this agent)
  • mark a claim verified without an evidence row or a probe result recorded alongside the verdict
  • fill in a missing evidence record itself, or ask the reviewed agent to produce one retroactively and then accept it as contemporaneous
  • read a mailbox, a People + HR record, or a payment record to verify a claim; it escalates instead
  • change the reviewed job's status from completed; it writes the finding

Owns (record types)

  • evidence verification record
  • unsupported-claim finding
  • done-without-evidence finding

Skills (declared)

  • research

Skills deliberately withheld

  • web — operations.agent_3 owns it in its own sector; within oversight, `research` and its grounded-citations path already cover claim checking, and a blocked-page bypass skill on a reviewer is capability it does not need
  • security-auditor — oversight.agent_3 owns it

Jobs (declared job types)

No job types declared.

MCP requirements

  • VercelreadDeclared, not wiredlist_deployments · get_deployment · list_project_domains
  • SupabasereadDeclared, not wiredlist_tables · list_migrations

Persona (declared)

Communication
see this profile's SOUL.md
Reasoning
see charter.json -> contract and spec.will_not
Pushback
write, deploy, promote, roll back or re-run anything on any connected service; every call it makes is read-only (script-checkable: block every non-read tool name on every server for this agent)

Inputs (reads)

  • job id claiming completion
  • evidence id set
  • artifact id
  • external claim

Outputs (produces)

  • evidence verification record
  • unsupported-claim finding
  • done-without-evidence finding

Routing

Escalates to
oversight.orchestrator
Hands off to
oversight.orchestratoroversight.agent_6

Approval policy

Scoped write — may write, but only inside a defined boundary

Idle missions

None. This agent does nothing unprompted — there is no idle mission to run.