Issue
Tool Rules and Prompt Rules Conflict
Tool, connector, function, or MCP rules conflict with prompt instructions, causing the AI or agent to face incompatible requirements.
What This Looks Like
The prompt tells the AI to do one thing, while the tool, connector, function, or MCP interface requires something incompatible. The agent may be instructed to format, omit, transform, call, or avoid something that the tool contract or tool rule requires differently.
Why It Matters
Tool use depends on both language instructions and interface rules. When they conflict, the AI may follow the prompt and fail the tool, follow the tool and appear to ignore the user, or produce an unstable compromise. Users may misread the failure as model disobedience when the rule set is incompatible.
Structural Signal
Two authority sources govern the same action or output with incompatible requirements. The issue is not only a bad tool call; it is a conflict between prompt-level rules and tool-level rules.
Common Triggers
- The prompt asks for a shape the tool schema does not accept
- Tool-required fields contradict prompt exclusions
- The tool supports fewer modes than the prompt assumes
- User instructions override values the tool requires
- MCP tool descriptions and prompt examples describe different contracts
- The agent is not given a precedence rule for prompt versus tool constraints
When to Use This Issue
Use this Issue when tool, connector, function, or MCP rules conflict with prompt instructions and the AI cannot satisfy both cleanly.
When Not to Use This Issue
Do not use this Issue when the tool call is merely malformed. Do not use it when the tool is unavailable or missing required inputs. This Issue applies when the declared rules themselves conflict.
Category
Primary Pattern
Declared Patterns
Authority Collision
A structural condition where multiple authorities claim governance over the same region without a declared precedence or resolution rule.
Interface Mismatch
A structural condition where observed interface behavior, shape, or exchange differs from the declared interface contract.
Authority Shadowing
A structural condition where a declared authority is functionally overridden by another authority without an explicit override rule.
Derived Primary Lenses
Authority Overlay Lens
Maps declared authority hierarchies onto observed structure to detect absence, override, or conflict.
Conflict Lens
Detects mutually incompatible constraints, claims, states, or declarations that cannot be simultaneously satisfied.
Interface Contract Lens
Compares declared interface structure to observed runtime structure to detect contract deviations.
Derived Secondary Lenses
Invariant Lens
Verifies that declared invariants are structurally enforceable and not contradicted by observed structure.
Normalization Lens
Transforms structurally equivalent variants into a canonical form to prevent false divergence.
Reconciliation Lens
Evaluates whether structural changes align with declared authority updates, version changes, or reconciliation rules.
Related AI-Adjacent Issues
Search Intents
- tool rules and prompt rules conflict
- MCP tool rules conflict with prompt
- tool schema conflicts with instruction
- AI prompt conflicts with tool requirement
- function rules contradict prompt
- agent tool instructions conflict