- Block REST API endpoints for label creation at the bash level for all agents.
- Restrict `forgejo_create_label` and related MCP tools for all agents.
- Restrict `forgejo_add_issue_labels` to only the `forgejo-label-manager`.
- Ensure all label operations are centralized through the `forgejo-label-manager`.
- Update agent definitions to use the label manager instead of direct API calls or MCP tools for adding labels.
This prevents agents from creating new project-level labels and enforces the use of organization-level labels, resolving the issue of duplicate labels being created.
Add explicit clone isolation protocols and warnings to prevent agents
from manipulating the local repository in /app. This ensures:
- Agents use isolated /tmp/ clones for all source code operations
- No interference between parallel agents
- No disruption to developer's local work environment
- No conflicts from branch changes or file modifications
Updated agents:
- Core implementation agents (implementer, build, plan)
- Quality gate agents (lint-fixer, typecheck-fixer, test-fixer, etc.)
- Test writing agents (behave-tester, unit-test-runner, coverage-improver)
- Analysis agents (difficulty-evaluator, fix-pr)
- Special cases (build-opencode with .opencode/ exception)
Each agent now includes prominent warnings and proper isolation protocols
with detailed explanations of why clone isolation is critical for
system stability.
- Add implementer-haiku.md as ultra-fast, cost-effective first tier
- Add behave-tester-haiku.md and robot-tester-haiku.md for fast testing
- Update difficulty-evaluator.md to assess 4 tiers (haiku/codex/sonnet/opus)
- Update subtask-loop.md with 4-tier escalation logic and state persistence
- Add PR comment state tracking for escalation recovery after restarts
- Conservative evaluator defaults to Haiku when in doubt for cost efficiency
- New escalation path: haiku → haiku → codex → sonnet → opus (forever)
- Includes state persistence to resume from correct tier after agent restarts