The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill defines a configuration format (.v8-project.json) for storing database and repository credentials. The SKILL.md file contains an example configuration with a hardcoded plaintext password ("password": "123"). While the documentation recommends using .gitignore, storing secrets in an unencrypted JSON file is a high-risk practice for credential exposure.
[INDIRECT_PROMPT_INJECTION]: The skill relies on reading the .v8-project.json file from the project directory, which presents a surface for indirect prompt injection if the file is modified by an untrusted source.
Ingestion points: The agent is instructed to use the Read tool to parse .v8-project.json from the local workspace.
Boundary markers: The skill does not define boundary markers or explicit instructions to treat the loaded JSON values as untrusted data.
Capability inventory: The skill possesses Read, Write, Glob, and AskUserQuestion tools. It prepares connection strings that are intended for shell execution by other skills.
Sanitization: There is no evidence of sanitization or validation of the data retrieved from the configuration file before it is used to generate commands or prompts.
[COMMAND_EXECUTION]: The skill provides templates for forming shell command arguments using data from the configuration file, such as /F "<path>" and /N"<user>" /P"<password>". This pattern is vulnerable to command injection if the configuration values contain shell metacharacters or are intentionally malformed by an attacker who has modified the .v8-project.json file.

db-list