Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/XxYouDeaDPunKxX/A.D.A.M.-Adaptive-Depth-and-Mode/llms.txt

Use this file to discover all available pages before exploring further.

SYS STATUS is a session health probe that measures whether A.D.A.M. is standing (core invariants holding) and functional (usable behavior intact) on the current host, at this moment. Unlike ADAM SELF TEST, which checks the full behavioral contract against a suite of discrete test cases, SYS STATUS runs six targeted micro-tests in a live scored table and returns a single summary line showing uptime and functional health.
SYS STATUS is an operator command. It is not required for normal operation — it is a diagnostic you run when you want a quick scored health check of the current session.

Trigger

Send the exact message:
SYS STATUS
A.D.A.M. runs the SYS_STATUS.md probe file. If that file is unavailable in visible context, it outputs exactly:
SOURCE_FILE_UNAVAILABLE SYS_STATUS.md
If the probe file is available but the strict output format cannot be satisfied, A.D.A.M. outputs exactly:
ADAM_UNSUPPORTED
SYS STATUS is available only while ACTIVE. In TRANSPORT or OFF, it is gated by the standard transport and off responses.

Output format

The output begins with the current mode tag (the usual first line for normal replies), followed by a single Markdown table and one final summary line. No questions, no extra commentary. The table has the exact header:
| Area | Micro-test | Expected | Observed (produced now) | Score (0-3) |
|---|---|---|---|---|
Each row scores a specific area from 0 to 3. The Observed cell shows what the model actually produces for that micro-test scenario — this is live output, not a description.

Final line

UPTIME: <0-3> | FUNCTIONAL: <0-3> | SCORES: TAG=<0-3>,FACTOID=<0-3>,BANTER=<0-3>,DEEP_GATING=<0-3>,GATING_NO=<0-3>,NO_INERTIA=<0-3>
Where:
  • UPTIME = rounded average of (TAG, DEEP_GATING, GATING_NO)
  • FUNCTIONAL = rounded average of (FACTOID, BANTER, NO_INERTIA)

Scored areas

Each row in the table targets a specific behavioral invariant:
AreaWhat it checks
TAGDoes the first line of a normal reply carry the correct mode tag?
FACTOID+AUTO_LIGHTDoes a factoid question (capital of germany) route to LOW with a minimal answer and no extras?
BANTER+LIGHT_MINIMALDoes a banter turn route to LOW with a short answer and no follow-up question?
DEEP_GATINGDoes a structural DEEP candidate produce MODE: MID -> POSSIBLE DEEP with the correct gating shape, ending in Switch to DEEP? (yes/no)?
GATING_NOAfter a no reply to gating, does A.D.A.M. stay out of DEEP?
NO_INERTIAAfter a DEEP reply, does a factoid question correctly route to LOW?
UPTIME (averaged from TAG, DEEP_GATING, GATING_NO) reflects whether the core structural invariants are holding: mode tagging, gating detection, and gating control. A low UPTIME score indicates the protocol scaffold itself may be degraded.FUNCTIONAL (averaged from FACTOID, BANTER, NO_INERTIA) reflects whether the session is delivering usable behavior: correct auto-routing for light input, correct brevity discipline, and correct no-inertia recomputation. A low FUNCTIONAL score with a high UPTIME score can indicate that routing heuristics are drifting even while the output structure is intact.

Internal effect

In addition to the scored output, SYS STATUS has one internal effect that is not printed:
Re-applies the active A.D.A.M. spec as SSOT and resets the State Overlay to EXPLORE for the next turn only.
This means the turn immediately following SYS STATUS starts from a clean overlay state regardless of prior session depth. The effect applies to the next turn only and does not persist.

Probe unavailability

If SYS_STATUS.md is not mounted in visible local context, the strict fallback is:
SOURCE_FILE_UNAVAILABLE SYS_STATUS.md
This is the same pattern used by ADAM SELF TEST when its probe file is absent. Neither command fabricates results without its source file.

Build docs developers (and LLMs) love