
Security News
Re-Enabled GitHub Actions Expose Thousands of Repositories to Mini Shai-Hulud
Two compromised GitHub Actions were re-enabled with malicious tags intact, exposing thousands of downstream repositories to Mini Shai-Hulud.
@spiderbrain/read
Advanced tools
Read a repo's committed deterministic context layer (.spiderbrain/): blast radius, keystones, and structure, offline. MCP server for AI coding agents, giving answers with receipts. Cloud why-layer with a key.
Use a repo's committed SpiderBrain understanding, offline. When a repo carries a
.spiderbrain/ folder, this reads it and answers structural questions about the
codebase with zero setup: what a change reaches (blast radius), which files are
load-bearing (keystones), and how any file fits. No account, no SpiderBrain
install. MIT, zero dependencies.
As an MCP server for your AI coding agent (Claude Code, Cursor):
npx @spiderbrain/read mcp
Tools: sb_blast, sb_keystones, sb_map, sb_ask. Your agent can now ask what
a change reaches and what matters, from the committed graph, deterministically.
One-off from the terminal:
npx @spiderbrain/read blast src/server/health.ts # what a change here reaches
npx @spiderbrain/read keystones # the load-bearing files
npx @spiderbrain/read map src/auth/session.ts # what a file is + touches
.spiderbrain/ bytes):
structure, dependencies, blast radius, keystones. Same question, same repo,
same answer, byte for byte.SPIDERBRAIN_API_KEY): the why-layer (the decision behind a file),
always-fresh scores, and semantic search. Get a key at
https://spiderbrain.ai/dashboard?tab=keys.export SPIDERBRAIN_API_KEY=sb_live_...
npx @spiderbrain/read why src/billing/charge.ts # the decision + reasoning
Every folder carries a fingerprint in manifest.json. The reader recomputes it
from structure.ndjson on load; a hand-edited or corrupted folder reports a
mismatch, so the map you query is the map that was published.
The public understanding set is source-free: file paths, structure, and an
edge-derivable blast radius only. Never your source code, never the scoring
weights. See the repo's .spiderbrain/manifest.json for the exact contents.
FAQs
Read a repo's committed deterministic context layer (.spiderbrain/): blast radius, keystones, and structure, offline. MCP server for AI coding agents, giving answers with receipts. Cloud why-layer with a key.
We found that @spiderbrain/read demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.

Security News
Two compromised GitHub Actions were re-enabled with malicious tags intact, exposing thousands of downstream repositories to Mini Shai-Hulud.

Research
/Security News
A malicious Firefox extension fetches its payload after installation to evade detection, steal Google session cookies, and automate account takeover.

Research
/Security News
The compromise affects MemTensor's MemOS, an open source memory framework for large language models (LLMs) and AI agents. Both npm package @memtensor/memos-cloud-openclaw-plugin and the PyPI package MemoryOS are compromised. They drop cross-platform Go binaries that exfiltrate developer secrets.