The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The scripts/detect-creds.sh utility accesses the user's local Sentry configuration file at ~/.sentryclirc to extract authentication tokens. This is used for a 'pre-flight' check to verify connectivity with the official Sentry API, which is a well-known service. While legitimate in the context of a Sentry management tool, it represents access to sensitive local credential stores.\n- [INDIRECT_PROMPT_INJECTION]: The skill implements a production debugging funnel in references/modes/mode-3-debug.md that ingests untrusted telemetry data, including error messages, stack traces, and logs from external sources via the Sentry CLI. \n
Ingestion points: Production error reports, breadcrumbs, and structured logs retrieved from Sentry and stored in temporary files like /tmp/iss.json.\n
Boundary markers: The skill does not define specific delimiters to isolate external telemetry content from the agent's core instructions.\n
Capability inventory: The agent is empowered to read local source code, execute diagnostic scripts, and run integration tests to verify fixes based on the ingested data.\n
Sanitization: While the skill provides robust mechanisms in references/privacy-redaction/ to redact data before it is sent to Sentry, it lacks equivalent sanitization for data retrieved from Sentry before it is processed by the agent.\n- [COMMAND_EXECUTION]: The skill provides several bash scripts (audit-project.sh, detect-creds.sh, triage-issues.sh, verify-tunnel.sh) that execute shell commands to perform repository analysis, network diagnostics, and CLI operations. These scripts rely on standard utilities such as grep, awk, curl, and jq.

use-sentry