The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses highly prescriptive and forceful language (e.g., "ABSOLUTELY MUST", "not negotiable", "not optional") and provides a table of "Red Flags" intended to override the agent's internal reasoning process. These patterns attempt to mandate the skill's workflow for all iOS/Swift-related tasks. Additionally, the skill introduces a surface for indirect prompt injection by ingesting and analyzing untrusted external data such as simulator console logs and crash reports.
Ingestion points: xclog captures simulator output; xcsym parses .ips, .crash, and .xccrashpoint files.
Boundary markers: None specified for the output generated by the tools or the content of the analyzed files.
Capability inventory: The skill executes xclog and xcsym CLI tools and performs local file writes to .axiom/preferences.yaml and .gitignore.
Sanitization: No explicit sanitization or escaping of ingested log/crash data is mentioned before it is processed by the agent.
[COMMAND_EXECUTION]: The skill provides detailed instructions for the agent to execute local CLI tools (xclog and xcsym) which are assumed to be distributed with the skill. These tools perform several operations that interact with the system and project files:
Process Management: xclog launch terminates and restarts targeted iOS applications to capture console output.
File Discovery: xcsym searches for dSYM files across multiple locations, including ~/Library/Developer/Xcode/Archives, ~/Library/Developer/Xcode/DerivedData, and ~/Downloads.
Project Modification: The skill instructs the agent to create a .axiom/ directory, write configuration to preferences.yaml, and automatically append entries to the project's .gitignore file.

axiom-tools