Install
$ agentstack add skill-amey-thakur-ai-skills-open-source-review-board ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Open source review board
An OSS review board is how a company keeps open source from becoming a legal and security liability it discovers too late. It fails when it is either a bottleneck everyone routes around or a rubber stamp that approves by default: dependencies arrive with incompatible licenses, employees push code to public repos with no policy, and a known-vulnerable package sits in the build because no one checked. The board's job is to make the safe path the easy path.
Method
- Gate every new dependency on license compatibility first. Maintain an
allow list and a deny list by license: permissive licenses like MIT, BSD, and Apache 2.0 usually clear; strong copyleft like GPL and AGPL needs review against how you distribute; anything unlicensed or with a custom license stops for a human. Automate the check in CI so the gate runs on every pull request, not on memory.
- Watch for license traps beyond the headline. Attribution requirements,
patent clauses, the AGPL network-use trigger, and the difference between linking and bundling. A permissive-looking package with a transitive GPL dependency is still a problem. The board reads the actual terms, not the badge on the README.
- Require a security review scaled to the dependency's reach. Scan with
something like an SCA tool or dependency scanner for known CVEs, then judge the project's health: maintenance activity, how fast it patches, single-maintainer risk, and how much of your attack surface it touches. A parser handling untrusted input earns a harder look than a build-time formatter.
- Set an explicit policy for contributing back. Decide when employees may
submit patches upstream, who signs the contributor license agreement, and what must be cleared before company code goes public. A clear yes-with-guardrails beats a vague no that engineers quietly ignore.
- Govern what you publish as its own decision. Releasing an internal project
as open source needs a license choice, a security scrub for secrets and internal references, a trademark check, and an owner for maintenance. Publishing is a commitment, not a giveaway; unowned public repos rot into liabilities.
- Keep an inventory and re-scan continuously. Maintain a software bill of
materials so you know what you depend on, and re-run scans as new CVEs land. A dependency that was clean at adoption can become the next incident; approval is a moment, exposure is ongoing.
Signals
- Does license checking run automatically in CI, or rely on someone remembering?
- Is the security review proportional to what the dependency can touch?
- Can you produce a current inventory of every OSS package you ship?
Boundaries
This governs adoption, contribution, and publication policy; it does not write the code or decide the architecture that needed the dependency. License interpretation at the edges is a legal judgment: escalate genuinely ambiguous or high-stakes terms to counsel rather than settling them at the board.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: Amey-Thakur
- Source: Amey-Thakur/AI-SKILLS
- License: MIT
- Homepage: https://amey-thakur.github.io/AI-SKILLS/
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.