The Agent Skills Directory

[SAFE]: The skill is a legitimate memory management extension for the Iskron platform. It utilizes vendor-specific tools (iskron_*) to manage data routing and persistence of user-relevant facts across different session contexts.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a potential attack surface because it retrieves and processes data from external project graphs (realms) which could be influenced by untrusted parties.
Ingestion points: Data is ingested from various realms via iskron_search and iskron_orient calls as described in the §1 Вспоминание (Recall) section of SKILL.md.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands when the agent reads content from these external graphs.
Capability inventory: The skill can perform write operations to the personal graph using iskron_update and manage user identity through iskron_me and iskronify integrations.
Sanitization: No explicit sanitization or validation protocols are mentioned for the content retrieved from external sources before it is used to influence the agent's behavior.
Assessment: This surface is inherent to the functionality of a semantic memory system, and no evidence of malicious exploitation was found in the provided files.
[SAFE]: The skill implements proactive security guidance by specifically instructing the agent to never store credentials, tokens, or private keys within the graph, as detailed in the 'Правило хранения' (Storage Rule) section.

minding