Instruction Anchor Guard
Analysis
This instruction-only skill is purpose-aligned, but it intentionally stores persistent instruction anchors that can influence future planning.
Findings (2)
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.
Checks for instructions or behavior that redirect the agent, misuse tools, execute unexpected code, cascade across systems, exploit user trust, or continue outside the intended task.
Source: unknown; Homepage: none; Version: 1.0.0
The registry metadata lacks a source/homepage and does not align with the SKILL.md frontmatter version `1.1.0`; this is a provenance/versioning note, not evidence of malicious behavior.
Checks for exposed credentials, poisoned memory or context, unclear communication boundaries, or sensitive data that could leave the user's control.
Store anchors in first available path: `memory/anchors.md` ... `.anchors.md`; ... Load active anchors ... Inject `ANCHOR_SET` into planning phase before tool execution
The skill intentionally persists user instructions and reuses them in future planning, which is purpose-aligned but means stale or mistaken anchors can keep influencing behavior.
