registry / @fingerskier/augment / 0.3.0

@fingerskier/augment@0.3.0

Local RAG memory system for coding agents

AI Security Review

scanned 11h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Explicit installation configures first-party Codex/Claude MCP and lifecycle hooks. Those hooks can start a local daemon, inject local memory context, and delay an agent stop once per turn.

Static reason

No blocking static signals were detected.; previous stored version diff introduced dangerous source

Trigger

User runs `augment install codex`, `augment install claude`, or `augment install all`.

Impact

Configured coding-agent sessions execute package-owned hooks and MCP tooling, gaining access to the configured local memory store.

Mechanism

Agent configuration and lifecycle-hook installation with a locally provisioned npm runtime.

Rationale

No malicious install-time behavior, credential theft, remote exfiltration, or arbitrary payload execution was found. Per policy, explicit first-party coding-agent integration setup with lifecycle hooks warrants a warning rather than a block.

Evidence

package.jsondist/cli.jsdist/install.jsdist/hooks.jsdist/daemon/http.jsdist/daemon/client.js~/.augment/npm-exec/node_modules/@fingerskier/augment/dist/bin/augment.js~/.agents/plugins/marketplace.json~/.claude.json~/.claude/CLAUDE.mdplugins/augment/.mcp.jsonplugins/augment/hooks/hooks.json.mcp.jsonCLAUDE.md

@fingerskier/augment@0.3.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code