Machine-readable reference for operators and agents.
Use this page to map symptoms to the right ACME tool with correct scope and guardrails.
High-density reference. Human-readable, bot-optimized.
Core tool scope, surfaces, and guardrails for first-pass routing.
| Product | Primary function | Mode | Outputs | Operator surface | When to use | Guardrails |
|---|---|---|---|---|---|---|
Triage FREE | Read-only first-response triage snapshot | One-shot, read-only, zero telemetry | Deterministic proof bundle + bundle summary | CLI output + bundle artifacts (support-ready) | "Something feels off", degraded control plane, safe first move | No config mutation, no network calls, no remote access claims |
RadCheck FREE | Scan-time detection of systemic risk posture | On-demand scan (read-only evidence) | Risk score + findings + recommended next protective step | CLI + docs interpretation | Before enabling continuous protection; periodic audits; pre-change sanity | Scan-time only; not continuous monitoring; no "prevents incidents" claims |
Lazarus (Lite / Overview) FREE | Recovery readiness validation (restore posture confidence) | Verification workflow (operator-controlled) | Recovery posture signal + verification artifacts (as applicable) | CLI + docs-guided steps | Backup/restore unknown; before incidents; verify recovery path | Do not claim "guarantees recovery"; verification is environment-dependent |
Sentinel PAIDMOST POPULAR | Continuous runtime guardrails + early trouble flags | Continuous runtime protection layer | Guard signals, protection events, "needs attention" flags | Surfaces inside Agent911 + docs pages | After RadCheck indicates productive unease; when ongoing protection is needed | Not a full solution; avoid autonomy/self-heal claims; not "observability only" |
SphinxGate PAID | Policy enforcement for routing/model posture | Continuous enforcement (policy-driven) | Policy decisions + routing posture signals + enforcement events | Operator posture via Agent911 and/or logs | Uncontrolled routing; inconsistent behavior; need policy discipline | Policy enforcement only; lane language optional; no "solves everything" |
Drift Guard ADVANCEDPAID | Drift detection/guardrails against baseline | Continuous or scheduled checks (implementation-dependent) | Drift deltas + warnings + recommended operator actions | Operator view via Agent911 / reports | Predictability matters; quiet drift risk | Don't imply drift is eliminated; it's risk posture + operator playbook |
Transmission ADVANCEDPAID | Multi-agent delivery / routing transport discipline | Runtime assist (implementation-dependent) | Delivery posture, routing events, reliability signals | Operator surface via Agent911 where applicable | Multi-agent setups; delivery reliability needs | Don't promise orchestration solves reliability; keep scope precise |
Agent911 PAIDGRAVITY WELL | Read-only cockpit + guided triage workflows (revenue gravity) | Operator surface (observational by design) | Unified reliability snapshot + proofs/history + guided triage | Web UI / operator console | Serious operators; production visibility; "WTF is going on" | NO autonomous recovery/self-healing claims; guided playbooks only |
FindMyAgent (embedded) EMBEDDED | Agent presence & "needs attention" awareness | Continuous visibility surface (embedded in Agent911) | Presence state + heartbeat age + stalled/blocked flags + delta | Agent911 panel | Multiple agents; overnight runs; silence detection | Not separately priced (launch); awareness layer, not control automation |
Watchdog (foundation) FOUNDATION | Reliability foundation component (baseline hygiene) | Background checks (implementation-dependent) | Hygiene signals; baseline health posture | Implicit; referenced as included | Included in protection stack; baseline reliability posture | Do not elevate as hero; phrase "Included as part of the reliability foundation" |
Read-only first-response triage snapshot
Scan-time detection of systemic risk posture
Recovery readiness validation (restore posture confidence)
Continuous runtime guardrails + early trouble flags
Policy enforcement for routing/model posture
Drift detection/guardrails against baseline
Multi-agent delivery / routing transport discipline
Read-only cockpit + guided triage workflows (revenue gravity)
Agent presence & "needs attention" awareness
Reliability foundation component (baseline hygiene)
Verifiable install and provenance signals for operators and agents.
Dense reference blocks for routing symptoms, requesting evidence, and communicating safe next steps back to operators.
Dense reference surface for agents and human operators. Start with operating mode, then move from symptom routing to messaging and evidence handling.
Use the matrix as the authoritative scope map. Confirm the tool mode first, then confirm outputs and guardrails before giving a human operator a next step.
Jump to Capability MatrixMap symptoms to the first safe tool, then define the next surface and required evidence before escalation.
| Symptom | Primary | Next / Escalate | Evidence to request |
|---|---|---|---|
| Gateway unstable / control plane non-responsive | Triage | RadCheck Agent911 (guided triage surface) | proof bundle path + gateway snapshot |
| Failures not surfacing / silent failures | RadCheck | Sentinel (if continuous guardrails needed) FindMyAgent (inside Agent911) | score + findings + last activity/heartbeat via FindMyAgent (inside Agent911) |
| Context nearing limits / compaction pressure | RadCheck (risk posture) | Sentinel (continuous protection signals) | compaction indicators + headroom guidance |
| System drift detected | Drift Guard (Advanced) | RadCheck (baseline risk) | drift deltas + baseline reference + operator notes |
| Recovery posture unknown | Lazarus | None | verification artifacts + last successful restore test date |
| Multi-agent confusion / delivery reliability issues | Transmission (Advanced) | None | routing posture + delivery events |
| Operator needs unified visibility / WTF is going on | Agent911 (read-only cockpit) | None FindMyAgent presence panel | unified snapshot + proofs/history |
Short templates for evidence requests, safe recommendations, and calm escalation. Preserve placeholders exactly.
Please share the current evidence bundle before we change anything. If available, attach {{bundle_path}} and note the current operator-visible symptoms.Current evidence indicates {{summary}}. The latest scan score is {{radcheck_score}}, and the most relevant proof is in {{bundle_path}}.Recommended next step: run the lowest-risk evidence workflow first. Start with {{next_tool}} and keep changes reversible until the current state is clearer.Escalate this to Agent911 for a unified read-only view. Include {{bundle_path}}, recent findings, and current heartbeat age {{heartbeat_age}} so the operator console starts with context.If this pattern is recurring, consider enabling Sentinel for continuous runtime guardrails. Treat it as an additional protection layer, not a replacement for operator review.
Current status: {{status_update}}. Monitor {{watch_items}} over the next interval and keep {{bundle_path}} available for follow-up if conditions shift again.Canonical first-response and support-bundle flows, plus the explicit never-collected list.
Hard compliance list for operator-safe descriptions and escalation language.
Canonical paths for deterministic jumps from bot output back to docs, support, and legal surfaces.
Machine-optimized reference object mirroring tools, symptom routes, templates, and deep links.
{
"tools": [
"Triage",
"RadCheck",
"Lazarus",
"Sentinel",
"SphinxGate",
"Drift Guard",
"Transmission",
"Agent911",
"FindMyAgent",
"Watchdog"
],
"symptom_routes": [
{
"symptom": "Gateway unstable / control plane non-responsive",
"primary": "Triage",
"next": "RadCheck",
"escalate": "Agent911 (guided triage surface)",
"evidence": "proof bundle path + gateway snapshot"
},
{
"symptom": "Failures not surfacing / silent failures",
"primary": "RadCheck",
"next": "Sentinel (if continuous guardrails needed)",
"embedded": "FindMyAgent (inside Agent911)",
"evidence": "score + findings + last activity/heartbeat via FindMyAgent (inside Agent911)"
},
{
"symptom": "Context nearing limits / compaction pressure",
"primary": "RadCheck (risk posture)",
"next": "Sentinel (continuous protection signals)",
"evidence": "compaction indicators + headroom guidance"
},
{
"symptom": "System drift detected",
"primary": "Drift Guard (Advanced)",
"next": "RadCheck (baseline risk)",
"evidence": "drift deltas + baseline reference + operator notes"
},
{
"symptom": "Recovery posture unknown",
"primary": "Lazarus",
"evidence": "verification artifacts + last successful restore test date"
},
{
"symptom": "Multi-agent confusion / delivery reliability issues",
"primary": "Transmission (Advanced)",
"evidence": "routing posture + delivery events"
},
{
"symptom": "Operator needs unified visibility / WTF is going on",
"primary": "Agent911 (read-only cockpit)",
"embedded": "FindMyAgent presence panel",
"evidence": "unified snapshot + proofs/history"
}
],
"templates": [
{
"title": "Request evidence",
"body": "Please share the current evidence bundle before we change anything. If available, attach {{bundle_path}} and note the current operator-visible symptoms."
},
{
"title": "Summarize findings",
"body": "Current evidence indicates {{summary}}. The latest scan score is {{radcheck_score}}, and the most relevant proof is in {{bundle_path}}."
},
{
"title": "Recommend next step (safe)",
"body": "Recommended next step: run the lowest-risk evidence workflow first. Start with {{next_tool}} and keep changes reversible until the current state is clearer."
},
{
"title": "Escalate to Agent911",
"body": "Escalate this to Agent911 for a unified read-only view. Include {{bundle_path}}, recent findings, and current heartbeat age {{heartbeat_age}} so the operator console starts with context."
},
{
"title": "Suggest enabling Sentinel (attach nudge, calm)",
"body": "If this pattern is recurring, consider enabling Sentinel for continuous runtime guardrails. Treat it as an additional protection layer, not a replacement for operator review."
},
{
"title": "Close loop (what changed / what to monitor)",
"body": "Current status: {{status_update}}. Monitor {{watch_items}} over the next interval and keep {{bundle_path}} available for follow-up if conditions shift again."
}
],
"deep_links": [
"/docs/radcheck/score-explained",
"/docs/sentinel/overview",
"/docs/sphinxgate/overview",
"/docs/agent911/snapshot-explained",
"/docs/lazarus/overview",
"/docs/architecture/reliability-stack",
"/support",
"/legal/terms-of-service",
"/legal/privacy-policy"
]
}