TokenOps maintainer agentplane claims the released tool enforces one token budget across an entire agent run before every call, potentially giving long-running agent harnesses a practical control against token-budget overruns.
state: expiredheat: lowuncertainty: highconvergesscott: mediumagent-harnesses inference-economicsagentplaneTokenOps
What is this?
TokenOps, maintained under theagentplane on GitHub and announced by agentplane on Show HN, describes itself as a control plane and SDK for run-scoped token governance in agent workflows. Its repository says LLM and tool calls share a run ID and ledger, with policies that can halt, mutate, or inject before the next call executes, including across processes. The supplied snippets support the advertised architecture, not a verified guarantee against overruns; the repository explicitly warns that embedded mode gives each process its own ledger and full budget, defeating the shared cross-process budget.
Why it matters to Scott
TokenOps offers a concrete implementation candidate for Scott’s Autonomy Budget and deterministic agent control plane, particularly Synthetic Futures Factory’s budget-scoped runs and concurrent worker lanes; this creates a practical evaluation opportunity rather than merely echoing his governance position. The claimed run-wide pre-call enforcement remains unverified, and embedded mode’s separate per-process budgets make cross-process enforcement a critical test before adoption; the supplied radar hits track related controls, not this TokenOps development.
ip:concept.autonomy-budgetdev:concept.deterministic-agent-control-planedev:project.synthetic-futuresradar:grapharc-runtime-agent-graph-gatesradar:decant-agent-token-accountingradar:concept.inference-economics
queries asked of Scott's wikis
- agent harness run-wide token budgets cost enforcement
- multi-agent shared state ledgers cross-process coordination
- pre-call policy gates tool execution governance
- long-running agents runaway loops termination controls
- budget-aware orchestration model routing task completion tradeoffs
Measured heat
no measured readings yet — the hourly heat pass fills this in
How the heat travelled
no chain yet — the hourly chain pass fills this in
Evidence (2) — ⭐ canonical anchor
Interpretation history
2026-09-09T08:29:42Z
Repeated checks have produced no evidence beyond the maintainer’s launch claim, and no concrete follow-up is expected; retire active monitoring without treating the budget-enforcement hypothesis as disproved. TokenOps remains an evaluation candidate, with shared-ledger concurrency and in-flight budget reservation the decisive tests if implementation evidence arrives.
2026-09-07T08:26:18Z
The staleness check adds no substantive evidence: TokenOps remains a relevant evaluation candidate, not a demonstrated run-wide budget guarantee, and the GitHub echo is not independent corroboration. Further attention should depend on implementation evidence or tests of shared-ledger concurrency and in-flight budget reservation, rather than repeated launch coverage.
2026-09-05T08:25:36Z
No new evidence changes TokenOps from a maintainer-claimed evaluation candidate into a demonstrated run-wide budget control; the GitHub echo repeats the announcement rather than independently corroborating it. The embedded-mode limitation recorded in grounding keeps shared-ledger concurrency and in-flight budget accounting central to any adoption test.
2026-09-05T08:24:41Z
grounded: converges/medium — TokenOps offers a concrete implementation candidate for Scott’s Autonomy Budget and deterministic agent control plane, particularly Synthetic Futures Factory’s
2026-09-05T08:22:14Z
case created — A linked first-party artifact presents a bounded run-level budget-enforcement claim distinct from existing cases, but evidence remains limited to the launch announcement.
Decision trace
- 09-09 18:29expireRepeated checks have produced no evidence beyond the maintainer’s launch claim, and no concrete follow-up is expected; retire active monitoring without treating the budget-enforcement hypothesis as di
- 09-09 18:29alert_silentThis delta contains only a staleness trigger, not a new release, implementation result, or actionable change. There is nothing new that Scott needs before the next briefing.
- 09-09 18:29alert_routeThis delta contains only a staleness trigger, not a new release, implementation result, or actionable change. There is nothing new that Scott needs before the next briefing.
- 09-07 18:26repriceThe staleness check adds no substantive evidence: TokenOps remains a relevant evaluation candidate, not a demonstrated run-wide budget guarantee, and the GitHub echo is not independent corroboration.
- 09-07 18:26alert_silentThere is no new release, access change, implementation result, or consequential participant in this delta. The previously identified evaluation opportunity can wait for a briefing; the staleness trigg
- 09-07 18:26alert_routeThere is no new release, access change, implementation result, or consequential participant in this delta. The previously identified evaluation opportunity can wait for a briefing; the staleness trigg
- 09-05 18:25repriceNo new evidence changes TokenOps from a maintainer-claimed evaluation candidate into a demonstrated run-wide budget control; the GitHub echo repeats the announcement rather than independently corrobor
- 09-05 18:25alert_silentThis reobservation adds no consequential delta to the previously assessed launch. The tool remains relevant to Scott’s budget-scoped agent runs, but evaluation can wait for the next briefing; no new i
- 09-05 18:25alert_routeThis reobservation adds no consequential delta to the previously assessed launch. The tool remains relevant to Scott’s budget-scoped agent runs, but evaluation can wait for the next briefing; no new i
- 09-05 18:24alert_silentThe first-party announcement establishes a new tool evaluation opportunity relevant to Scott’s budget-scoped agent runs, but does not validate its enforcement guarantees. The supplied evidence offers
- 09-05 18:24surface_candidateThe first-party announcement establishes a new tool evaluation opportunity relevant to Scott’s budget-scoped agent runs, but does not validate its enforcement guarantees. The supplied evidence offers
- 09-05 18:24alert_routeThe first-party announcement establishes a new tool evaluation opportunity relevant to Scott’s budget-scoped agent runs, but does not validate its enforcement guarantees. The supplied evidence offers
- 09-05 18:24groundTokenOps offers a concrete implementation candidate for Scott’s Autonomy Budget and deterministic agent control plane, particularly Synthetic Futures Factory’s budget-scoped runs and concurrent worker
- 09-05 18:22createA linked first-party artifact presents a bounded run-level budget-enforcement claim distinct from existing cases, but evidence remains limited to the launch announcement.