Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Python agent-contextagent-memoryagent-pluginsagentic-ragcontext-databasedsh-pluginknowledge-basememorymemory-managementself-evolvingvector-database
"""Public-safe diagnostics for OpenClaw subprocess failures."""
from __future__ import annotations
import re
MAX_DIAGNOSTIC_CHARS = 800
_SECRET_PATTERNS = (
(re.compile(r"(?i)\b(bearer)\s+\S+"), r"\1 <redacted>"),
(
re.compile(
r"(?i)\b(api[_-]?key|authorization|access[_-]?token|token|secret)"
r"(\s*[:=]\s*)([^\s,;]+)"
),
r"\1\2<redacted>",
),
(re.compile(r"\bsk-[A-Za-z0-9_-]{8,}\b"), "<redacted>"),
)
def compact_subprocess_diagnostic(value: str, limit: int = MAX_DIAGNOSTIC_CHARS) -> str:
"""Redact common credential shapes and bound subprocess output for CI logs."""
compact = " ".join((value or "").split())
for pattern, replacement in _SECRET_PATTERNS:
compact = pattern.sub(replacement, compact)
if not compact:
return "<empty>"
if len(compact) <= limit:
return compact
return f"{compact[:limit]}...<truncated>"
def format_openclaw_cli_failure(message: str, stderr: str, returncode: int) -> str:
"""Build a bounded error without copying raw OpenClaw stderr."""
diagnostic = compact_subprocess_diagnostic(stderr)
return f"{message} (exit={returncode}; stderr={diagnostic})"