cleveragents/cleveragents-core

Fork 3

BUG-HUNT: [error-handling] GraphExecutor._initialize_context() hardcodes paper-writing stage names creating incorrect defaults for non-paper-writing workflows #7348

New issue

Open

opened 2026-04-10 17:58:40 +00:00 by HAL9000 · 3 comments

HAL9000 commented

2026-04-10 17:58:40 +00:00

Owner

Bug Report: [error-handling] GraphExecutor._initialize_context() hardcodes paper-writing stage names polluting context for non-paper-writing reactive applications

Severity Assessment

Impact: All reactive graph executions get a context pre-populated with paper-writing stages (intro, discovery, brainstorming, paper_review, latex_generation, etc.) regardless of the actual use case, which could confuse routing logic or cause unexpected agent invocations
Likelihood: High — affects all reactive graph executions that don't explicitly override these context keys
Priority: Medium

Location

File: src/cleveragents/reactive/graph_executor.py
Function/Class: GraphExecutor._initialize_context()
Lines: ~90-125

Description

The _initialize_context() method is supposed to initialize the graph execution context with "required default keys." However, it hard-codes a set of stages that are specific to a paper-writing workflow (intro, discovery, brainstorming, vetting, structure, section_writing, paper_review, latex_generation).

This has several problems:

Incorrect defaults: A generic reactive agent running a code review, customer service, or data processing pipeline would get a context with paper-writing stages injected, which is semantically wrong
Stage validation bug: The condition requires_refresh checks if "section_writing", "paper_review", and "latex_generation" are in stage_order — if they're not (as in any non-paper-writing use case), it forces the paper-writing default stages onto the context
writing_stage key: Every context gets a writing_stage key set to "intro", even for workflows that have nothing to do with writing
paper_details key: Every context gets a paper_details dict with topic, length, audience, publication, format, other, even for non-paper workflows

Evidence

def _initialize_context(self) -> dict[str, Any]:
    """Ensure global context has required default keys for graph execution."""
    context = self._config.global_context if self._config else {}
    default_stage_order = [
        "intro",
        "discovery",
        "brainstorming",
        "vetting",
        "structure",
        "section_writing",
        "paper_review",
        "latex_generation",   # <-- Paper writing specific!
    ]
    stage_order = context.get("stage_order")
    requires_refresh = not isinstance(stage_order, list) or any(
        stage not in (stage_order or [])
        for stage in ("section_writing", "paper_review", "latex_generation")  # Paper stages!
    )
    if requires_refresh:
        context["stage_order"] = list(default_stage_order)  # Forces paper stages!
    if "writing_stage" not in context or context.get("writing_stage") not in (
        context.get("stage_order") or []
    ):
        context["writing_stage"] = "intro"  # Paper writing state!
    if "paper_details" not in context:
        context["paper_details"] = {  # Paper-specific fields injected everywhere!
            "topic": None,
            "length": None,
            "audience": None,
            "publication": None,
            "format": None,
            "other": None,
        }
    return context

Expected Behavior

The _initialize_context() method should only inject defaults that are universally required for graph execution (e.g., routing state). Paper-writing specific stage names and the paper_details dict should be defined in the paper-writing actor/route configuration, not in the generic graph executor.

Actual Behavior

Every reactive graph execution context is pre-populated with paper-writing workflow data. A code review bot or customer service agent would get writing_stage = "intro" and paper_details = {topic: None, ...} injected into its context, which could:

Confuse routing rules that check context["writing_stage"]
Waste memory with irrelevant context data
Break routing logic if the agent's rules coincidentally use the same key names with different semantics

TDD Note

After this bug issue is verified, a corresponding Type/Testing issue will be created for TDD. The test will use tags: @tdd_issue, @tdd_issue_, and @tdd_expected_fail to prove the bug exists before fixing it.

Automated by CleverAgents Bot
Supervisor: Bug Detection Pool | Agent: bug-hunt-pool-supervisor

## Bug Report: [error-handling] GraphExecutor._initialize_context() hardcodes paper-writing stage names polluting context for non-paper-writing reactive applications ### Severity Assessment - **Impact**: All reactive graph executions get a context pre-populated with paper-writing stages (`intro`, `discovery`, `brainstorming`, `paper_review`, `latex_generation`, etc.) regardless of the actual use case, which could confuse routing logic or cause unexpected agent invocations - **Likelihood**: High — affects all reactive graph executions that don't explicitly override these context keys - **Priority**: Medium ### Location - **File**: `src/cleveragents/reactive/graph_executor.py` - **Function/Class**: `GraphExecutor._initialize_context()` - **Lines**: ~90-125 ### Description The `_initialize_context()` method is supposed to initialize the graph execution context with "required default keys." However, it hard-codes a set of stages that are **specific to a paper-writing workflow** (`intro`, `discovery`, `brainstorming`, `vetting`, `structure`, `section_writing`, `paper_review`, `latex_generation`). This has several problems: 1. **Incorrect defaults**: A generic reactive agent running a code review, customer service, or data processing pipeline would get a context with paper-writing stages injected, which is semantically wrong 2. **Stage validation bug**: The condition `requires_refresh` checks if `"section_writing"`, `"paper_review"`, and `"latex_generation"` are in `stage_order` — if they're not (as in any non-paper-writing use case), it **forces** the paper-writing default stages onto the context 3. **`writing_stage` key**: Every context gets a `writing_stage` key set to `"intro"`, even for workflows that have nothing to do with writing 4. **`paper_details` key**: Every context gets a `paper_details` dict with `topic`, `length`, `audience`, `publication`, `format`, `other`, even for non-paper workflows ### Evidence ```python def _initialize_context(self) -> dict[str, Any]: """Ensure global context has required default keys for graph execution.""" context = self._config.global_context if self._config else {} default_stage_order = [ "intro", "discovery", "brainstorming", "vetting", "structure", "section_writing", "paper_review", "latex_generation", # <-- Paper writing specific! ] stage_order = context.get("stage_order") requires_refresh = not isinstance(stage_order, list) or any( stage not in (stage_order or []) for stage in ("section_writing", "paper_review", "latex_generation") # Paper stages! ) if requires_refresh: context["stage_order"] = list(default_stage_order) # Forces paper stages! if "writing_stage" not in context or context.get("writing_stage") not in ( context.get("stage_order") or [] ): context["writing_stage"] = "intro" # Paper writing state! if "paper_details" not in context: context["paper_details"] = { # Paper-specific fields injected everywhere! "topic": None, "length": None, "audience": None, "publication": None, "format": None, "other": None, } return context ``` ### Expected Behavior The `_initialize_context()` method should only inject defaults that are **universally required** for graph execution (e.g., routing state). Paper-writing specific stage names and the `paper_details` dict should be defined in the paper-writing actor/route configuration, not in the generic graph executor. ### Actual Behavior Every reactive graph execution context is pre-populated with paper-writing workflow data. A code review bot or customer service agent would get `writing_stage = "intro"` and `paper_details = {topic: None, ...}` injected into its context, which could: 1. Confuse routing rules that check `context["writing_stage"]` 2. Waste memory with irrelevant context data 3. Break routing logic if the agent's rules coincidentally use the same key names with different semantics ### Category error-handling ### TDD Note After this bug issue is verified, a corresponding Type/Testing issue will be created for TDD. The test will use tags: @tdd_issue, @tdd_issue_<this-issue-number>, and @tdd_expected_fail to prove the bug exists before fixing it. --- **Automated by CleverAgents Bot** Supervisor: Bug Detection Pool | Agent: bug-hunt-pool-supervisor

HAL9000 referenced this issue

2026-04-10 18:08:49 +00:00

[AUTO-BUG-SUP] Bug Hunt Status (Cycle 1) #7340

HAL9000 added the

State

Unverified

label

2026-04-10 18:37:43 +00:00

freemo added the

labels

2026-04-12 18:35:52 +00:00

HAL9000 added

and removed

labels

2026-04-14 06:24:12 +00:00

HAL9000 commented

2026-04-14 06:24:12 +00:00

Author

Owner

✅ Verified — Bug: GraphExecutor hardcodes paper-writing stage names creating incorrect defaults. MoSCoW: Should-have. Priority: Medium.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor

✅ **Verified** — Bug: GraphExecutor hardcodes paper-writing stage names creating incorrect defaults. MoSCoW: Should-have. Priority: Medium. --- **Automated by CleverAgents Bot** Supervisor: Project Owner | Agent: project-owner-pool-supervisor

HAL9000 commented

2026-04-14 06:29:50 +00:00

Author

Owner

✅ Verified — Bug: GraphExecutor hardcodes paper-writing stage names creating incorrect defaults. MoSCoW: Should-have. Priority: Medium.

Automated by CleverAgents Bot
Supervisor: Project Owner | Agent: project-owner-pool-supervisor