- Add mandatory PROJECT-LEVEL code checks (lsp_diagnostics at src/ or . level)
- Strengthen verification duties with explicit QA checklist
- Add 'SUBAGENTS LIE - VERIFY EVERYTHING' reminders throughout
- Emphasize that only orchestrator sees full picture of cross-file impacts
- Remove 'Start Sisyphus work session' text check, keep only <session-context> tag
- Update interactive_bash description with WARNING: TMUX ONLY emphasis
- Update tests to use <session-context> wrapper
OpenCode API returns different structures for user vs assistant messages:
- User: info.model = { providerID, modelID } (nested)
- Assistant: info.modelID, info.providerID (flat top-level)
Previous code only checked nested format, causing model info loss when
continuation hooks fired after assistant messages.
Files modified:
- todo-continuation-enforcer.ts
- ralph-loop/index.ts
- sisyphus-task/tools.ts
- background-agent/manager.ts
Added test for assistant message model extraction.
- Unified verification (Step 1-3) and post-verification (Step 4-6) into continuous workflow
- Step 4: Immediate plan file marking after verification passes
- Step 5: Commit atomic unit
- Step 6: Proceed to next task
- Emphasized immediacy: 'RIGHT NOW - Do not delay'
- Applied to both boulder state and standalone reminder contexts
- Move Turn Termination Rules inside <system-reminder> block (from line 488 to ~186)
- Add Final Constraint Reminder at end of prompt (constraint sandwich pattern)
- Preserve all existing interview mode detail and strategies
Applies OpenCode's effective constraint patterns to prevent plan-mode agents
from offering to implement work instead of staying in consultation mode.
- Change skills type from string[] to string[] | null
- Empty array [] now returns error with available skills list
- null is allowed for tasks that genuinely need no skills
- Updated tests to use skills: null instead of skills: []
- Forces explicit decision: either specify skills or justify with null
The prometheus-md-only hook was not enforcing file restrictions because
getSessionAgent() returned undefined - setSessionAgent was only called
in message.updated event which doesn't always provide agent info.
- Add setSessionAgent call in chat.message hook when input.agent exists
- Add session state tests for setSessionAgent/getSessionAgent
- Add clearSessionAgent cleanup to prometheus-md-only tests
This ensures prometheus-md-only hook can reliably identify Prometheus
sessions and enforce .sisyphus/*.md write restrictions.
When transitioning from Prometheus (Planner) to Sisyphus via /start-work,
the session agent was not being cleared. This caused prometheus-md-only
hook to continue injecting READ-ONLY constraints into sisyphus_task calls.
- Add clearSessionAgent() call when start-work command is detected
- Add TDD test verifying clearSessionAgent is called with sessionID
- Add shared/system-directive.ts with SYSTEM_DIRECTIVE_PREFIX constant
- Unify all system message prefixes to [SYSTEM DIRECTIVE: OH-MY-OPENCODE - ...]
- Add isSystemDirective() filter to keyword-detector to skip system messages
- Update prometheus-md-only tests to use new prefix constants
- Delete src/hooks/preemptive-compaction/ entirely
- Remove dcp_for_compaction from schema and executor
- Clean up related imports, options, and test code
- Update READMEs to remove experimental options docs
The tools restriction was defined in sisyphus.ts but not enforced in
config-handler.ts like other agents (orchestrator-sisyphus, Prometheus).
Added explicit tools setting to guarantee call_omo_agent is disabled.
Remove OpenCode built-in tools (lsp_goto_definition, lsp_find_references, lsp_symbols, lsp_servers) that are not provided by oh-my-opencode. Keep only lsp_diagnostics, lsp_prepare_rename, lsp_rename.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
lsp_goto_definition, lsp_find_references, lsp_symbols are provided by OpenCode, not oh-my-opencode. Keep only the 4 tools we actually provide: lsp_diagnostics, lsp_servers, lsp_prepare_rename, lsp_rename.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
The watermark (commit footer and co-author) was inconsistently applied because:
1. The skill tool didn't receive gitMasterConfig
2. The approach was 'default ON, inject DISABLED override' which LLMs sometimes ignored
This refactors to 'inject only when enabled' approach:
- Remove hardcoded watermark section from base templates
- Dynamically inject section 5.5 based on config values
- Default is still ON (both true when no config)
- When both disabled, no injection occurs (clean prompt)
Also fixes missing config propagation to skill tool and createBuiltinAgents.
feat(concurrency): prevent background task races and leaks
Summary
Fixes race conditions and memory leaks in the background task system that could cause "all tasks complete" notifications to never fire, leaving parent sessions waiting indefinitely.
Why This Change?
When background tasks are tracked for completion notifications, the system maintains a pendingByParent map to know when all tasks for a parent session are done. Several edge cases caused "stale entries" to accumulate in this map:
1. Re-registering completed tasks added them back to pending tracking, but they'd never complete again
2. Changing a task's parent session left orphan entries in the old parent's tracking set
3. Concurrent task operations could cause double-acquisition of concurrency slots
These bugs meant the system would sometimes wait forever for tasks that were already done.
What Changed
- Concurrency management: Added proper acquire/release lifecycle with cleanup on process exit (SIGINT, SIGTERM)
- Parent session tracking: Fixed cleanup order. Now clears old parent's tracking before updating parent ID
- Stale entry prevention: Only tracks tasks that are actually running; actively cleans up completed tasks
- Renamed registerExternalTask → trackTask: Clearer name (the old name implied external API consumers, but it's internal)
Update BackgroundManager to rename the method for tracking external tasks, improving clarity and consistency in task management. Adjust related tests to reflect the new method name.
Previously, continuation hooks (todo-continuation, boulder-continuation, ralph-loop)
and background tasks resolved model info from filesystem cache, which could be stale
or missing. This caused session.prompt to fallback to default model (Sonnet) instead
of using the originally configured model (e.g., Opus).
Now all session.prompt calls first try API (session.messages) to get current model
info, with filesystem as fallback if API fails.
Affected files:
- todo-continuation-enforcer.ts
- sisyphus-orchestrator/index.ts
- ralph-loop/index.ts
- background-agent/manager.ts
- sisyphus-task/tools.ts
- hook-message-injector/index.ts (export ToolPermission type)
- Delete PSModulePath in beforeEach() to prevent CI cross-platform detection
- Set SHELL=/bin/bash to ensure tests start with clean Unix-like environment
- Fixes flaky test failures on GitHub Actions CI runners
- Tests can still override these values for PowerShell-specific behavior
Add formatDetailedError helper that includes:
- Full args dump (description, category, agent, skills)
- Session ID and agent info
- Stack trace (first 10 lines)
Applied to all catch blocks for better debugging.
When user types '/start-work my-plan ultrawork', the hook now:
1. Extracts plan name from <user-request> section
2. Strips ultrawork/ulw keywords from the plan name
3. Searches for matching plan (exact then partial match)
4. Uses the matched plan instead of resuming stale boulder state
This fixes the bug where '/start-work [PLAN] ultrawork' would:
- Include 'ultrawork' as part of the plan name argument
- Ignore the explicit plan and resume an old stale plan from boulder.json
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
