sensitive-logging-audit

por openai

Audita y corrige la exposición de datos sensibles a través del registro de Python en openai-agents-python. Úsalo al revisar logging, print, warnings, stderr, traceback,…

npx skills add https://github.com/openai/openai-agents-python --skill sensitive-logging-audit

Sensitive Logging Audit

Objective

Find candidate output sinks, trace their values manually, fix demonstrated leaks at shared runtime boundaries, and prove redaction with adversarial tests.

The collector is only a syntax-based search aid. It does not resolve Python aliases or control flow, certify policy guards, or prove that an absent candidate is safe.

Workflow

1. Establish the review surface

  • Work in the current checkout and preserve unrelated changes.
  • Read src/agents/_debug.py, src/agents/logger.py, and the affected callers.
  • Treat exception messages, arguments, tracebacks, causes, contexts, notes, names, URLs, and arbitrary values as potentially sensitive.
  • Read the Python redaction validation matrix.

Run the collector tests, then collect candidates:

uv run python .agents/skills/sensitive-logging-audit/scripts/test_inventory.py
uv run python .agents/skills/sensitive-logging-audit/scripts/inventory_logging.py \
  --format json --output /tmp/sensitive-logging-candidates.json

The report intentionally contains no policy, safe, or guard classification.

2. Supplement the collector with source search

The collector does not follow assignments such as emit = logger.error. Search the source directly and inspect aliases, callbacks, wrappers, and reflective dispatch:

rg -n '\.(debug|info|warning|warn|error|exception|critical|fatal|log)\b' src/agents
rg -n '\b(print|pprint|pp|warn|warn_explicit|write|writelines|print_exc|print_exception)\b' src/agents
rg -n 'DONT_LOG_(MODEL|TOOL)_DATA|log_(model|tool|model_and_tool)_action' src/agents

Do not turn collector coverage or a textual guard into a security conclusion. Trace producers and callers.

3. Classify manually

Assign each reviewed path one disposition:

  • model: model requests, responses, Realtime events, or derived values.
  • tool: tool arguments, outputs, MCP data, tool events, or derived values.
  • model+tool: either class may reach the sink.
  • operational: demonstrated to contain only non-sensitive SDK metadata.
  • intentional-output: explicitly user-facing output rather than diagnostics.
  • uncertain: source tracing is incomplete.

Record evidence in the audit report. The script does not validate or inherit dispositions.

4. Fix runtime boundaries

Before changing runtime behavior, use $implementation-strategy.

  • Check the relevant _debug.DONT_LOG_MODEL_DATA and _debug.DONT_LOG_TOOL_DATA flags before formatting or inspecting sensitive values.
  • Redact mixed model/tool values when either flag disables data logging.
  • In redacted mode, emit a fixed message and omit sensitive args, extra, and exc_info.
  • Build diagnostic-only context lazily so redacted mode never reads it.
  • Preserve useful diagnostics when sensitive-data logging is explicitly enabled.
  • Keep logging failure from changing fallback, cleanup, event, rejection, or cancellation behavior.
  • For MCP URLs, remove credentials, query parameters, and fragments in diagnostic mode; never use sanitized names as a substitute for fixed redacted messages.

5. Prove caller behavior

Add tests at every changed caller boundary. Inspect the complete LogRecord, not only rendered text. Test both redacted policies, diagnostic mode, hostile objects, exception chains, and the caller's observable fallback or cleanup behavior as applicable.

6. Re-run and close out

Re-run the collector, the manual searches, focused tests, and applicable repository gates. Use $code-change-verification for runtime or test changes and $pr-draft-summary when required.

Report candidate counts as search coverage only. Lead with confirmed leaks fixed, retained intentional output, reviewed uncertainty, and verification results. Never report a clean collector result as proof that no sensitive logging path exists.

Más skills de openai

release
openai
Realizar un lanzamiento de Symphony incrementando la versión confirmada, integrándola, etiquetando el commit fusionado y verificando el flujo de trabajo de lanzamiento de Burrito. Usar cuando se solicite…
signing-entitlements
openai
Inspeccionar problemas de firma, entitlements, runtime reforzado y Gatekeeper en aplicaciones de macOS. Usar cuando se solicite diagnosticar fallos de firma de código, entitlements faltantes,…
building-ai-agent-on-cloudflare
openai
Crea agentes de IA en Cloudflare usando el Agents SDK con gestión de estado, WebSockets en tiempo real, tareas programadas, integración de herramientas y chat…
epigraphdb-skill
openai
Envía solicitudes compactas a la API de EpiGraphDB para ontología, literatura, MR, genes-fármacos y evidencia de vías de soporte. Úsalo cuando un usuario desee resúmenes concisos de EpiGraphDB.
runtime-behavior-probe
openai
Planificar y ejecutar investigaciones de comportamiento en tiempo de ejecución con scripts de sonda temporales, matrices de validación, controles de estado e informes centrados en hallazgos. Usar solo cuando…
deep-security-scan
openai
Úsalo cuando el usuario solicite un escaneo de seguridad de Codex profundo, exhaustivo, de múltiples pasadas o que reduzca la varianza, en todo el repositorio o en una ruta específica. Ejecuta pasadas independientes repetidas…
define-security-policy
openai
Definir, revisar o actualizar la guía de SECURITY.md para un repositorio o componente. Úselo cuando el usuario quiera aclarar qué debe revisar Codex Security, qué está fuera de…
validation
openai
Usar cuando Codex ya está en la fase de validación de un escaneo de seguridad o el usuario solicita explícitamente determinar si uno o más posibles hallazgos de seguridad…