Non-Claude models skip planning and under-parallelize. Two new sections
injected only when model is not Claude:
- Plan Agent Dependency: multi-step tasks MUST consult Plan Agent first,
use session_id for follow-ups, ask aggressively when ambiguous
- Deep Parallel Delegation (rewrite): explicit '4 units = 4 agents'
pattern, each with clear GOAL + success criteria, all run_in_background
Sisyphus prompt instructed 'your next action is background_output' which
caused agents to repeatedly poll running tasks instead of ending their
response and waiting for the system notification.
- Replace 'STOP all other output' with 'end your response' (actionable)
- Add system-reminder notification mechanism explanation
- Add explicit 'Do NOT poll' prohibition
- Reduce background_cancel(all=true) mentions from 5x to 1x (Hard Blocks)
- Reduce Oracle collect obligation from 4x to 2x
- Remove motivational fluff ('blind spots', 'normal and expected')
Net: -2 lines, clearer mechanism, eliminates polling loop root cause.
The auto-commit section instructed Hephaestus to automatically commit after
implementation work. Users who didn't know about this behavior would get
surprise commits — a trust-breaking behavioral change flagged by 5 Oracle
reviews as the sole publish blocker for 3.9.0.
date is already injected by OpenCode's system.ts. omo-env now contains only
Timezone and Locale, which are stable across requests and never break cache.
Current time with HH:MM:SS changed every second, invalidating the prompt cache
on every request. Date-level precision is sufficient; timezone and locale are
stable. Removes Current time field entirely from createEnvContext output.
The test 'hephaestus is created when github-copilot provider is connected'
had incorrect expectation. github-copilot does not provide gpt-5.3-codex,
so hephaestus should NOT be created when only github-copilot is connected.
This test was causing CI flakiness due to incorrect assertion and
missing readConnectedProvidersCache mock (state pollution between tests).
Also adds cacheSpy mock for proper isolation.
- fix(hooks): skip todo continuation when agent has pending question (#1888)
Add pending-question-detection module that walks messages backwards
to detect unanswered question tool_use, preventing CONTINUATION_PROMPT
injection while awaiting user response.
- fix(config): allow custom agent names in disabled_agents (#1693)
Change disabled_agents schema from BuiltinAgentNameSchema to z.string()
and add filterDisabledAgents helper in agent-config-handler to filter
user, project, and plugin agents with case-insensitive matching.
- fix(agents): change primary agents mode to 'all' (#1891)
Update Sisyphus, Hephaestus, and Atlas agent modes from 'primary'
to 'all' so they are available for @mention routing and task()
delegation in addition to direct chat.
Counter Gemini's tendency to skip Phase 0 intent classification by
injecting a mandatory self-check gate before tool calls. Includes
intent type classification, anti-skip mechanism, and common mistake
table showing wrong vs correct behavior per intent type.
🤖 Generated with assistance of [OhMyOpenCode](https://github.com/code-yeongyu/oh-my-opencode)
Add 'gemini' to prompt source types and route Gemini models to new
Gemini-optimized prompts via isGeminiModel detection. Update barrel
exports for all 3 agent modules. All existing tests pass.
Add buildDeepParallelSection() function that injects guidance for non-Claude
models on parallel deep agent delegation:
- Detect when model is non-Claude and 'deep' category is available
- Inject instructions to decompose tasks and delegate to deep agents in parallel
- Give goals, not step-by-step instructions to deep agents
- Update Sisyphus prompt builder to pass model and call new function
This helps GPT-based Sisyphus instances leverage deep agents more effectively
for complex implementation tasks.
🤖 Generated with assistance of OhMyOpenCode
Revert test name and assertion to original behavior per PR review feedback.
The test now correctly expects Atlas to respect uiSelectedModel instead of using its own fallback chain.
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
transformModelForProvider only handled github-copilot provider, leaving
google provider models untransformed. This caused ProviderModelNotFoundError
when google/gemini-3-flash was sent to the API (correct ID is
gemini-3-flash-preview).
Changes:
- Add google provider to transformModelForProvider with idempotent regex
negative lookahead to prevent double -preview suffix
- Fix category-default path in model-resolution-pipeline when
availableModels is empty but connected provider exists
- Fix getFirstFallbackModel first-run path that constructed raw model IDs
without transformation
- Fix github-copilot provider gemini transforms to also use idempotent
regex (was vulnerable to double-transform)
- Extract transformModelForProvider to shared module (single source of
truth, imported by cli and shared layers)
- Add 20 new test cases: unit tests for both providers, runtime
integration tests for category-default and fallback-chain paths,
double-transform prevention for both providers
- Added a new configuration option `disable_omo_env` to control the injection of the `<omo-env>` block in agent prompts.
- Updated relevant functions and tests to support this feature, ensuring that the environment context can be toggled on or off as needed.
- Enhanced documentation to reflect the new option and its implications for API cost and cache hit rates.
- Remove formatCustomSkillsBlock function (dead code)
- Remove unused truncateDescription import
- Update buildCategorySkillsDelegationGuide to compact format
- Update tests to match new compact output
Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-opencode)
Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
Add Step 0 intent extraction to counter GPT 5.2's conservative grounding bias:
- Map surface questions to true action intent (e.g., "Did you do X?" → do X now)
- Verbalization pattern: model must state intent before acting, creating commitment
- Turn-end self-check to prevent stopping after only talking about work
Prevents Hephaestus from answering questions then stopping when action is implied.
Atlas and Sisyphus prompts instructed agents to use background_cancel(all=true)
before final answers. This destroys uncollected background task results and
contradicts existing NEVER directives in the Sisyphus prompt, causing agents
to lose explore/librarian outputs mid-session.
Replace with individual task cancellation pattern that preserves completed
task results while still cleaning up running disposable tasks.
YeonGyu-Kim