ARCHITECT: Autonomous Goal Execution for AI Agents
Security checks across malware telemetry and agentic risk
Overview
This instruction-only skill is coherent with its advertised purpose, but users should understand that it deliberately makes the agent more autonomous and may use persistent memory if paired with other skills.
Before installing, be comfortable with an agent that plans and executes multi-step goals after approval. Check each mission brief, insist on confirmation for external or irreversible actions, and review any memory/reflection behavior if you use this with a memory skill.
VirusTotal
66/66 vendors flagged this skill as clean.
Risk analysis
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.
The agent may take more initiative in planning and completing tasks after a goal is approved.
The skill intentionally instructs the agent to perform multi-step autonomous execution. This is purpose-aligned, and the same artifact includes confirmation rules, but users should understand it may use available tools without step-by-step prompting.
When you receive a high-level goal, run this loop autonomously... EXECUTE... VALIDATE... ADAPT
Review the mission brief before approving, and require explicit confirmation for any action that writes, sends, deletes, publishes, or uses external systems.
The agent may reframe vague requests and pursue an interpretation the user did not originally state.
The skill tells the agent to reinterpret the user's stated request into a 'real goal.' This is disclosed and mitigated by the mission-brief confirmation step, but users should check that the inferred goal matches their intent.
The stated goal is rarely the real goal. Before decomposing, extract... REAL GOAL
Read the mission brief carefully and correct the goal, constraints, and out-of-scope items before typing YES.
If memory features are enabled, task details or inferred preferences could be reused later.
The skill describes saving lessons for future runs and also promotes pairing with a memory skill. This is purpose-aligned, but persistent context can carry sensitive or stale information into later tasks.
8. REFLECT → Log insights for future executions
Avoid storing secrets or private data in reflections, and periodically review or clear any memory entries created during use.
Users have less external provenance information for the skill publisher and project history.
The registry metadata does not provide a source repository or homepage. The skill is instruction-only with no code to install, so this is a provenance note rather than a behavior concern.
Source: unknown; Homepage: none
Install only if you trust the registry listing and review the included instructions before enabling the skill.