Skip to content

Latest commit

 

History

History
123 lines (77 loc) · 6.68 KB

File metadata and controls

123 lines (77 loc) · 6.68 KB

Evidence-driven program priority

program/registry.json records research priorities. Opsle Tasks is the current workload and task-management authority. Edit the registry, not this generated view.

Operating question

How can current Opsle Tasks workloads use less intelligence and context while preserving correctness and defensible evidence?

Anti-nitpick guardrail

Do not create a work item solely because an implementation can be improved.

Normally admit new work only for:

  • violated invariant
  • demonstrated defect
  • measured inefficiency
  • missing capability blocking the current program objective
  • experiment requirement
  • security or safety issue
  • externally required release condition

Park by default:

  • cosmetic cleanup
  • architectural taste
  • hypothetical robustness
  • speculative future requirement

NOW / NEXT / THEN / LATER / PARKED

Lane Objective Repositories Entry gate
NOW Maintain current Tasks workload and reconcile demonstrated defects with shipped recovery evidence; Notebook implementation and production integration are complete. tasks, research An authorized Tasks work item and a qualifying evidence-backed reason are required.
NEXT Address demonstrated integration or measurement deficiencies; shipped capability packages alone do not establish current-runtime compatibility or research completion. gearbox, context-firewall, visible-value, affected-verification, decision-evidence-protocol, agent-trajectory-profiler An authorized Tasks work item and a qualifying evidence-backed reason are required.
THEN Investigate semantic edits only when workload evidence establishes a need. semantic-edit-protocol An authorized Tasks work item and a qualifying evidence-backed reason are required.
LATER Prepare controlled research and public material when evidence and separate release authority support it. site An authorized Tasks work item and a qualifying evidence-backed reason are required.
PARKED Retain organization housekeeping separately from workload eligibility. .github An authorized Tasks work item and a qualifying evidence-backed reason are required.

NOW — Maintain current Tasks workload and reconcile demonstrated defects with shipped recovery evidence; Notebook implementation and production integration are complete.

Entry: An authorized Tasks work item and a qualifying evidence-backed reason are required.

Exit: The scoped defect or question is resolved with bounded evidence; release requirements remain separate.

NEXT — Address demonstrated integration or measurement deficiencies; shipped capability packages alone do not establish current-runtime compatibility or research completion.

Entry: An authorized Tasks work item and a qualifying evidence-backed reason are required.

Exit: The scoped defect or question is resolved with bounded evidence; release requirements remain separate.

THEN — Investigate semantic edits only when workload evidence establishes a need.

Entry: An authorized Tasks work item and a qualifying evidence-backed reason are required.

Exit: The scoped defect or question is resolved with bounded evidence; release requirements remain separate.

LATER — Prepare controlled research and public material when evidence and separate release authority support it.

Entry: An authorized Tasks work item and a qualifying evidence-backed reason are required.

Exit: The scoped defect or question is resolved with bounded evidence; release requirements remain separate.

PARKED — Retain organization housekeeping separately from workload eligibility.

Entry: An authorized Tasks work item and a qualifying evidence-backed reason are required.

Exit: The scoped defect or question is resolved with bounded evidence; release requirements remain separate.

Opsle Tasks boundary

Opsle Tasks is the current workload and task-management authority. Its repository is opsle/tasks.

Measure: Gearbox, Context Firewall, Visible Value, Affected Verification.

Without separate authorization, do not:

  • public release
  • DNS or TLS changes
  • launch provider work

Evidence-triggered concept activation

  • routing → gearbox
  • durable state → tasks
  • retry or deterministic preflight → gearbox
  • context overload → context-firewall
  • verification selection → affected-verification
  • recovery → tasks

Visible Value target

Baseline: The configured primary model and reasoning effort performs the same task work and receives raw unfiltered evidence.

Savings require an inspectable same-work baseline or a clearly labeled estimate derived from that baseline; marketing counterfactuals are prohibited.

  • MEASURED — The value comes directly from deterministic artifacts or provider usage records and still carries the applicable Visible Value evidence class, normally EXACT or OBSERVED.
  • DERIVED — The value is a reproducible calculation over identified measured inputs and records its method and applicable Visible Value evidence class.
  • ESTIMATED — The value carries the ESTIMATED evidence class and names its method, assumptions, pricing source, and uncertainty.
  • UNAVAILABLE — The value is not currently supported and must remain absent rather than being rendered as zero.

Per-child receipt: child/task identity; model; reasoning effort; Gearbox route; routing rationale; input tokens; output tokens; raw evidence/context size; Context Firewall retained size; reduction percentage; estimated tokens avoided; estimated cost avoided; duration; attempt number; success/failure; escalation/retry reason; retry potentially avoidable through deterministic preflight.

Task/run summary: total children; model/effort distribution; total model tokens consumed; work completed deterministically without model use; Context Firewall reduction; estimated token/cost savings; first-pass success rate; repair-child rate; tokens spent on retries; avoidable-intelligence estimate.

Later

  • controlled empirical experiments
  • frozen real-workload benchmark corpus
  • independent replication
  • Opsle Tasks public and self-hosted release
  • opsle.com research and public site
  • hosted Opsle offering

Parked

  • Cosmetic cleanup, architectural polishing, hypothetical robustness and speculative requirements without qualifying evidence.

Exact next execution

Select only authorized, evidence-backed work from Opsle Tasks. Notebook implementation and production integration are complete through PR #83; the pending #52 scope is superseded by that recovery and is not a retry candidate. Existing lane membership is unchanged; see reconciliation evidence in program/registry.json.