The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from Salesforce debug logs which could theoretically contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill is triggered by and processes .log files from Salesforce orgs as defined in SKILL.md.
Boundary markers: Although the skill uses structured output templates, it lacks explicit safety instructions for the agent to ignore or delimit potentially malicious instructions embedded within the processed logs.
Capability inventory: The skill can execute shell commands via the Salesforce CLI and generate Apex code, which provides a high-capability environment that could be targeted by an injection attack.
Sanitization: There is no evidence of sanitization or character escaping applied to the log content before it is parsed by the agent.
[COMMAND_EXECUTION]: The skill requires the execution of system commands via the Salesforce CLI to perform its core functions of log retrieval and org management.
Evidence: references/cli-commands.md and references/analysis-playbook.md provide numerous commands for log retrieval and management, such as sf apex get log and sf data delete record.
Evidence: The skill includes bash script templates (save-latest-log.sh, monitor-errors.sh) that automate these CLI interactions for monitoring and cleanup tasks.
[DATA_EXFILTRATION]: The skill's primary function involves retrieving debug logs from a remote Salesforce environment to the local user context, which may contain sensitive application logic or business data.
Evidence: SKILL.md and references/cli-commands.md focus on downloading log data for analysis, which is a standard part of the Salesforce troubleshooting process.

debugging-apex-logs