Memory Fortress
v1.1.0Memory Fortress manages agent memory with prioritized, domain-organized files using compaction-aware saving, boot loading, scoring, promotion, and bloat prev...
⭐ 0· 76·0 current·0 all-time
byPeter Zsak@wroadd
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
The name/description (agent memory management) match the actual instructions: the SKILL.md only directs reading and writing of structured files (state/, memory/, MEMORY.md, archive) to implement boot, scoring, promotion/demotion, and compaction-aware saving. No unrelated binaries, env vars, or installs are requested.
Instruction Scope
Instructions stay within the stated purpose (file-based memory management). However, the runtime guidance explicitly tells agents to persist users' exact words and decisions to disk immediately, which can cause sensitive user data or secrets to be retained long-term. The SKILL.md also asserts agents 'execute this automatically' at boot — this is an instruction-level expectation but consistent with a memory-management skill; ensure the agent runtime actually grants file access only where intended.
Install Mechanism
No install spec and no code files — the skill is instruction-only. This minimizes risk from downloads or arbitrary code execution because nothing is written to disk by an installer.
Credentials
No environment variables, credentials, or external service tokens are requested. The file paths referenced are all scoped to agent memory (state/, memory/, archive/), which is proportionate to the described function.
Persistence & Privilege
The skill does not request 'always: true' and does not modify other skills; autonomous invocation is allowed (platform default). The notable privilege is its instruction to persist user-provided text immediately into workspace files, giving the skill persistent storage in the agent environment — this is expected for a memory manager but has privacy implications.
Assessment
This skill is internally consistent for agent memory management, but it will save users' exact words, decisions, and discovered facts to files in the agent workspace (state/, memory/, MEMORY.md, daily logs, archive). Before installing, consider: (1) whether you want user inputs or secrets stored persistently; (2) who can read the agent workspace and whether those files should be access-controlled or encrypted; (3) retention/archival policy (the skill recommends moving logs to archive after 14 days and keeping MEMORY.md <=10KB — confirm your agent will enforce that); and (4) whether you need redaction or filtering to prevent accidental storage of passwords/API keys. If any of these are concerns, do not enable the skill until you have file-permission controls and retention/redaction rules in place.Like a lobster shell, security has layers — review code before you run it.
latestvk97dg9q704v195rncgvz0gx3zs83deay
License
MIT-0
Free to use, modify, and redistribute. No attribution required.