D.E.E.P. Framework
ReviewAudited by ClawScan on May 10, 2026.
Overview
This is mostly a local personality-memory helper, but it advertises a critical safety check that is only mock logic and always approves actions.
Install only if you want a persistent local personality-memory framework. Keep sensitive information out of the memory files, review the generated soul_vault.json, and do not treat the included triple-check tool as a real safety approval mechanism.
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.
A user or agent could treat this as a real safety gate for important actions when it provides only a fixed approval response.
SKILL.md describes deep_triple_check as a mandatory safety/alignment filter before critical actions, but the actual implementation is explicitly mock logic and unconditionally approves the action.
# Mock logic for the triple-check filter ... "T1_CONSTITUTION": "PASSED" ... "verdict": "PROCEED"
Do not rely on deep_triple_check as a security or safety control. If installed, treat it as a placeholder until it performs real, user-reviewable checks and can reject unsafe actions.
The stored files may contain personal, relationship, goal, or behavioral information and could steer future agent responses if over-trusted.
The skill intentionally creates persistent local personality and relationship memory, which may later be reused by the agent as context.
Agents must maintain the following structures in `memory/personality/` ... RELATIONSHIPS: Trust metrics & social graph ... CORE MEMORIES: Narrative continuity.
Review and edit the memory/personality files regularly, avoid storing secrets, and make sure higher-priority user and system instructions override any saved personality content.