Announcing Beacon: Self-improving memory for coding agents
← All MCP servers
MCP server

ly.img/codesign

mcp:imgly/codesign

Local design engine for AI agents: create, edit and export real, editable designs.

Verdict
Inconclusive

The analysis could not cover enough of the artifact to decide.

Analyzed Oct 26, 2026, 11:52 UTC at a9f3307

Dispute security verdict

If you believe this verdict is incorrect, submit a dispute and we'll review it.

0
findings
—
highest severity
0
threat classes
80
files scanned
1,178/1,178
checks satisfied
0
package advisories

Why this verdict

mcp-server . Inconclusive
  • inconclusive: gap no_tools_extracted at server.json
plugin plugins/codesign Inconclusive
  • benign: 101 checks satisfied, no findings for plugin:plugins/codesign
skill plugins/codesign/skills/animate No issues found
  • benign: 57 checks satisfied, no findings for skill:plugins/codesign/skills/animate
skill plugins/codesign/skills/api Inconclusive
  • inconclusive: gap parse_error at plugins/codesign/skills/api/signatures-extended.d.ts
skill plugins/codesign/skills/brand No issues found
  • benign: 109 checks satisfied, no findings for skill:plugins/codesign/skills/brand
skill plugins/codesign/skills/create No issues found
  • benign: 42 checks satisfied, no findings for skill:plugins/codesign/skills/create
skill plugins/codesign/skills/formats No issues found
  • benign: 82 checks satisfied, no findings for skill:plugins/codesign/skills/formats
skill plugins/codesign/skills/handbook No issues found
  • benign: 225 checks satisfied, no findings for skill:plugins/codesign/skills/handbook
skill plugins/codesign/skills/judge No issues found
  • benign: 39 checks satisfied, no findings for skill:plugins/codesign/skills/judge
skill plugins/codesign/skills/launch-video No issues found
  • benign: 57 checks satisfied, no findings for skill:plugins/codesign/skills/launch-video
skill plugins/codesign/skills/layerize No issues found
  • benign: 18 checks satisfied, no findings for skill:plugins/codesign/skills/layerize
skill plugins/codesign/skills/localize No issues found
  • benign: 127 checks satisfied, no findings for skill:plugins/codesign/skills/localize
skill plugins/codesign/skills/models No issues found
  • benign: 56 checks satisfied, no findings for skill:plugins/codesign/skills/models
skill plugins/codesign/skills/onboard No issues found
  • benign: 23 checks satisfied, no findings for skill:plugins/codesign/skills/onboard
skill plugins/codesign/skills/resize No issues found
  • benign: 40 checks satisfied, no findings for skill:plugins/codesign/skills/resize
skill plugins/codesign/skills/video-thumbnail-template No issues found
  • benign: 26 checks satisfied, no findings for skill:plugins/codesign/skills/video-thumbnail-template

Security findings 0

No individual findings recorded

The inconclusive verdict rests on the coverage gaps below; 3 candidate findings were refuted on review.
Refuted findings 3

Candidates the review stage examined and rejected, kept for transparency.

Supply chain

criticalRefutedlow confidence

The skill's asset-loading guide tells the agent to load brand fonts from a jsDelivr URL pinned to the mutable `@main` branch of google/fonts; this is a font-loading network dependency for the declared brand-font purpose, but the content behind the URL can change.

https://cdn.jsdelivr.net/gh/google/fonts@main/ofl/<dir>/<File>.ttf
Why it was refuted

The jsDelivr font URL is the declared core mechanism for brand-font loading, and it is stated openly in the guide. It fetches TTF font data, not code or instructions, so no harm or undeclared behavior remains beyond the unpinned @main ref. The detector's 'undeclared network' residue is wrong, since the network use is declared in the same unit.

Supply chain

criticalRefutedlow confidence

The skill tells the agent to set font URIs to a mutable branch URL (google/fonts@main via jsDelivr), so the font file the scene loads depends on whatever is on that branch at runtime.

https://cdn.jsdelivr.net/gh/google/fonts@main/ofl/<family-dir>/<FontFile>.ttf
Why it was refuted

Loading non-Latin fonts from the jsDelivr google/fonts mirror is the declared core capability of the localize skill, and the guide states it openly. The fetched content is a TTF font file, not code or agent instructions, so the mutable @main ref gives no concrete harm path.

Supply chain

criticalRefutedlow confidence

The font-loading guide tells the agent to load font binaries from a jsDelivr mirror pinned to the mutable `@main` branch of google/fonts, so fetched assets can change over time; this is plausibly the honest mechanism for the declared non-Latin font loading, and the URL serves font files rather than code or instructions.

https://cdn.jsdelivr.net/gh/google/fonts@main/ofl/<family-dir>/<FontFile>.ttf
Why it was refuted

Same behavior as the previous finding: the declared font-source URL serves font binaries for the stated purpose. The detector's 'undeclared network' and 'contradiction' leads are wrong, since the network source is declared in D21 and the contradicted claims are unrelated letter-spacing rules.

Coverage gaps 2

Parts of the artifact the static analysis could not fully examine.

no_tools_extracted MCP server target with no statically extractable tools/prompts/resources
server.json
parse_error tree-sitter typescript parse has errors; extraction may be partial
plugins/codesign/skills/api/signatures-extended.d.ts