Context Manager

ReviewAudited by ClawScan on May 10, 2026.

Overview

This skill is a mostly coherent local memory/search manager, but its main implementation is loaded from an unprovided local workspace module, so the executable behavior is not fully reviewable.

Before installing, verify the missing context-manager implementation and dependencies from a trusted source. If you use it, avoid saving secrets, understand where local memory and logs are stored, and only configure OpenAI credentials when you are comfortable with the data that may be analyzed.

Findings (3)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

Medium

#ASI04: Agentic Supply Chain Vulnerabilities

What this means

If a context_manager module already exists or is later placed at that path, this skill may execute code that was not included in the registry review.

Why it was flagged

The visible skill wrapper loads its core implementation from a local workspace path that is not included in the provided file manifest and has no install spec, leaving the main executable behavior and provenance unclear.

Skill content

CONTEXT_MANAGER_DIR = Path.home() / ".openclaw-autoclaw" / "workspace" / "context-manager"
sys.path.insert(0, str(CONTEXT_MANAGER_DIR))
from context_manager import context_manager

Recommendation

Install only after obtaining and reviewing the actual context-manager implementation and dependencies from a trusted source; the package should include or pin the core implementation instead of relying on an unreviewed workspace path.

Low

#ASI03: Identity and Privilege Abuse

What this means

Using the optional LLM feature can consume the user's OpenAI quota and may send selected context to the OpenAI service depending on the unreviewed implementation.

Why it was flagged

The skill documents optional OpenAI API key use for LLM functionality, while registry metadata lists no required credentials or environment variables; this appears purpose-aligned but should be visible to users.

Skill content

export OPENAI_API_KEY="your-api-key"

Recommendation

Configure an OpenAI key only if you need LLM analysis, use a restricted key where possible, and avoid storing or sending secrets unless you understand the implementation.

Low

#ASI06: Memory and Context Poisoning

What this means

Information saved to this skill may remain searchable later, including across agents if cross-agent search is used.

Why it was flagged

The skill explicitly persists agent files, vectors, cache, and logs locally for later retrieval, which is expected for a long-term memory tool but can preserve sensitive context across tasks.

Skill content

所有数据存储在：
~/.openclaw-autoclaw/workspace/context-manager/
├── agents/
├── vector_db/
├── cache/
└── logs/

Recommendation

Do not store secrets unless necessary, periodically delete unneeded agents/files, and verify who or what can access the OpenClaw workspace directory.