Deterministic JavaScript Semantic Analysis for an Autonomous Agent

Outcome

Deterministic JavaScript Semantic Analysis for an Autonomous Agent

I need a deterministic, machine-readable representation of the supplied JavaScript module that can be consumed by another autonomous software agent before it modifies the module.

The completed result must determine, from the source itself:

• the module's semantic role,
• its risk level,
• observed filesystem capabilities,
• observed network capabilities,
• observed process or execution capabilities,
• observed security-sensitive capabilities,
• constraints that should be preserved during future evolution,
• operations that should be treated as forbidden,
• and behaviour that cannot be determined reliably from the source.

The result must explicitly distinguish between:

• capability not observed,
• capability observed,
• behaviour unknown or not determinable.

Unknown or undeterminable behaviour must never be silently treated as safe or absent.

The representation must be deterministic and include an integrity hash that another agent can independently verify from the canonical representation.

Do not modify or execute the supplied JavaScript source.

The final deliverable must be machine-readable and self-contained. A downstream autonomous code-evolution agent must be able to use it without access to the original source.

Done means

1. A complete machine-readable semantic representation is delivered.
2. All requested capability categories are explicitly represented.
3. Observed, not-observed, and undeterminable states are distinguishable.
4. Risk, constraints, and forbidden operations are explicit.
5. The representation is deterministic.
6. An integrity hash is included and independently reproducible.
7. The original source remains unmodified and is not executed.