9bf9fc295d
Add 6 shared reference docs that standardize agent behavior across all GSD workflows: - universal-anti-patterns.md: 27 behavioral rules across 8 categories - context-budget.md: 4-tier system (PEAK/GOOD/DEGRADING/POOR) with read-depth rules based on context window size - gate-prompts.md: 14 named prompt patterns for decision points - revision-loop.md: Check-Revise-Escalate with stall detection (max 3) - domain-probes.md: domain-specific follow-up questions for 10 domains - agent-contracts.md: completion markers for all 21 GSD agents Wire into existing workflows via @-reference includes: - plan-phase.md: revision-loop, gate-prompts, agent-contracts - execute-phase.md: agent-contracts, context-budget - discuss-phase.md: domain-probes, gate-prompts, universal-anti-patterns
3.3 KiB
Context Budget Rules
Standard rules for keeping orchestrator context lean. Reference this in workflows that spawn subagents or read significant content.
See also: references/universal-anti-patterns.md for the complete set of universal rules.
Universal Rules
Every workflow that spawns agents or reads significant content must follow these rules:
- Never read agent definition files (
agents/*.md) --subagent_typeauto-loads them - Never inline large files into subagent prompts -- tell agents to read files from disk instead
- Read depth scales with context window -- check
context_window_tokensin.planning/config.json:- At < 500000 tokens (default 200k): read only frontmatter, status fields, or summaries. Never read full SUMMARY.md, VERIFICATION.md, or RESEARCH.md bodies.
- At >= 500000 tokens (1M model): MAY read full subagent output bodies when the content is needed for inline presentation or decision-making. Still avoid unnecessary reads.
- Delegate heavy work to subagents -- the orchestrator routes, it doesn't execute
- Proactive warning: If you've already consumed significant context (large file reads, multiple subagent results), warn the user: "Context budget is getting heavy. Consider checkpointing progress."
Read Depth by Context Window
| Context Window | Subagent Output Reading | SUMMARY.md | VERIFICATION.md | PLAN.md (other phases) |
|---|---|---|---|---|
| < 500k (200k model) | Frontmatter only | Frontmatter only | Frontmatter only | Current phase only |
| >= 500k (1M model) | Full body permitted | Full body permitted | Full body permitted | Current phase only |
How to check: Read .planning/config.json and inspect context_window_tokens. If the field is absent, treat as 200k (conservative default).
Context Degradation Tiers
Monitor context usage and adjust behavior accordingly:
| Tier | Usage | Behavior |
|---|---|---|
| PEAK | 0-30% | Full operations. Read bodies, spawn multiple agents, inline results. |
| GOOD | 30-50% | Normal operations. Prefer frontmatter reads, delegate aggressively. |
| DEGRADING | 50-70% | Economize. Frontmatter-only reads, minimal inlining, warn user about budget. |
| POOR | 70%+ | Emergency mode. Checkpoint progress immediately. No new reads unless critical. |
Context Degradation Warning Signs
Quality degrades gradually before panic thresholds fire. Watch for these early signals:
- Silent partial completion -- agent claims task is done but implementation is incomplete. Self-check catches file existence but not semantic completeness. Always verify agent output meets the plan's must_haves, not just that files exist.
- Increasing vagueness -- agent starts using phrases like "appropriate handling" or "standard patterns" instead of specific code. This indicates context pressure even before budget warnings fire.
- Skipped steps -- agent omits protocol steps it would normally follow. If an agent's success criteria has 8 items but it only reports 5, suspect context pressure.
When delegating to agents, the orchestrator cannot verify semantic correctness of agent output -- only structural completeness. This is a fundamental limitation. Mitigate with must_haves.truths and spot-check verification.