The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The scripts/publish-signal.sh script relies on a PRIVATE_KEY environment variable for cryptographic signing. Passing raw private keys through environment variables is a high-risk practice that can lead to credential theft if the process environment is logged or inspected.
[PROMPT_INJECTION]: The skill presents multiple injection surfaces by instructing the agent to consume untrusted external content.
Ingestion points: SKILL.md instructs the agent to fetch and follow a remote instruction file (https://bankrsignals.com/heartbeat.md) and poll a public signal feed (https://bankrsignals.com/api/feed).
Boundary markers: Absent. There are no instructions provided to the agent to treat the remote markdown or signal reasoning fields as untrusted data.
Capability inventory: The agent can sign messages via the Bankr API or local scripts to publish trades and signals.
Sanitization: Absent. The agent processes instructions and trading signals directly from the provider's infrastructure.
[COMMAND_EXECUTION]: The scripts/publish-signal.sh script uses node -e to dynamically execute Javascript code for signing messages. This execution pattern is a risk vector if environment variables or arguments are manipulated, though it is currently used for internal logic.

bankr-signals