# memory/budgets.md - DOME v0.5

Budgets, transitions and lint (proposed starting values). It complements AGENTS.md and never contradicts it.

- This contract: target 3,000 tokens. If exceeded, move per-layer detail to linked documents.
- Queried index: 1,000 tokens. Wiki page: 1,500. Memory retrieved per task: 3,000.
- Measure with the model's tokenizer, including inherited instructions and installed skill descriptions. Promote a behavior to a skill only after three manual repetitions; archive unused skills.
- Warn at 80%. At 100%, split, compact or retrieve in stages. Never truncate constraints to meet a number.
- Compact while preserving decisions, dates, exceptions, contradictions and sources. Validate before replacing the index.
- Initial review_by: candidates 7 days, procedures 30, decisions 90 or earlier if their premise changes.
- After 90 days without use, review for archiving. Do not archive active constraints only because nobody queried them.
- Archive superseded or inactive entries and keep their references.
- Invalidate falsehoods with a reason and a link to the correction. Evicting from context does not delete evidence. Deleting requires applying retention.md to sources, derivatives and copies. Never promise deletion while data persists in history or backups.

Source of the skill maintenance rule: Collin Wilkins, collinwilkins.com/articles/building-production-ready-ai-agent-skills. Numeric values are DOME proposals.

Lint: after ingest and significant closes, validate metadata, links, index, duplicates, contradictions, permissions and budgets. Weekly, also review validity, orphans and retention. Always include directives and execution: untestable criteria, expired directives, untested tools and broken links between documents. Record finding, evidence, severity, owner and status. Review semantic changes and deletions according to permissions. Do not mark a finding resolved without checking it.