Skip to main content

Module types

Module types 

Source

Structs§

CommandInfo
The result of classifying a command — its effect, matched subcommand, escalation flags, affected paths, env gates, and wrapper info. Produced by classify.
CommandKnowledge
What we know about a command — its effect, subcommands, flags, env gates, and path semantics. Knowledge entries come from embedded defaults or user-provided TOML config; they describe commands without making policy decisions.
CommandProperties
FlagSchema
Flag-level knowledge for subcommand extraction and path/escalation detection.
KnowledgeBase
The full knowledge base — all known commands and wrappers. Populated from embedded defaults and extended by user/project TOML config.
PathSpec
Which arguments are filesystem paths — used by the policy layer for path-aware authorization.
SubcommandEntry
A subcommand’s knowledge: its effect, flags, env gates, path semantics, and optional nested subcommands for multi-level commands.
SubcommandMap
Map of subcommand patterns to their entries. Patterns are space-separated strings (e.g. "pr create") matched via longest_match.
WrapperInfo
Wrapper metadata returned when the base command is a known wrapper.
WrapperKnowledge
Semantic knowledge about a wrapper command, layered on top of agent-shell-parser’s WrapperSpec (which handles stripping mechanics).

Enums§

Effect
The effect level of a command or subcommand.
EnvGate
Environment variable conditions that modify classification.
PathPositionals
How positional arguments map to filesystem paths.

Constants§

MAX_SUBCOMMAND_DEPTH
Maximum number of words that can form a subcommand pattern.