Resume-effect reconciliation for checkpointed agent runs. From $9/mo.
Scored across 5 tools
Most tools are clearly distinct: audit journal completeness, execute a guarded effect, pin a discovery digest, record/backfill an effect, and produce a signed verdict. The closest overlap is between resume_execute_effect and resume_record_effect, and to a lesser extent resume_audit_journal versus resume_verdict, but the descriptions give enough context to distinguish them.
All tools share the resume_ prefix and use snake_case, with most following a verb_noun pattern (audit_journal, execute_effect, pin_discovery, record_effect). resume_verdict is noun-only rather than verb_noun, which is a minor deviation but still readable and consistent with the domain.
Five tools is well-scoped for a focused resume-effect reconciliation workflow. Each tool maps to a distinct step or artifact in that workflow, and there is no obvious padding.
The set covers auditing, executing, recording, pinning, and verdict generation, but there are notable gaps: no tool to inspect or list existing effects/ledger entries, no unpin or ledger correction operation, and no explicit run/checkpoint status retrieval. Agents may hit dead ends when they need to read state before acting.
Journal-completeness audit for a killed run (Team tier, POST /audit).
Inputs: run_id (required), journal (required), checkpoint_ns (required).
Guarded exactly-once execution against the demo target (POST /effects/execute).
Inputs: kind (required), node (required), run_id (required), effect_key (required), +2 more.
Pin a checkpoint discovery digest (POST /pin).
Inputs: run_id (required), checkpoint_ns (required), discovery_pin (required).
Record (backfill) an effect in your account's ledger (POST /effects).
Inputs: kind (required), node (required), run_id (required), effect_key (required), +3 more.
Signed resume verdict for a checkpointed run (subscriber token; same computation and account scoping as POST /verdict).
Inputs: node (required), run_id (required), checkpoint_ns (required), declared_effects (required), +1 more.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Agent checkpoints. Resume after context resets and handoffs with retry-safe, versioned saves.
Checkpoint and resume for AI agent runs, so crashed runs pick up where they left off.
Outcome-verified watchdog for scheduled agents. $0.002/check.
Hosted runtime for persistent agent teams, durable workflows, memory, schedules, and goals.
Add one secure layer between your agents and this server.
