You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit a5092dd
Browse filesBrowse the repository at this point in the historyBrowse files
Refresh the public SkillSpector tree from the internal OSS release branch
release/oss-2026-06-16 at 94599d2baae5d532d83f630ddeeeead1f33c8cbb.
This snapshot was generated by scripts/create-oss-release.sh, which removes
internal-only files and produces a single orphan-branch tree for public
publication.
Changes:
- Bump public package snapshot to 2.2.3.
- Carry Python metadata and uv.lock support for >=3.12,<3.15, including Python 3.14.
- Add native provider-backed LangChain chat model creation, including Anthropic support.
- Resolve Anthropic API call failures caused by unsupported maximum/minimum parameters in structured-output tool schemas.
- Add shared OpenAI-compatible chat model helper.
- Add live provider test targets, docs, and provider endpoint tests.
- Keep default unit and coverage targets scoped away from provider/integration tests.
- Update LLM response text handling and provider unit coverage.
Resolves#4, #34, #66, #76.
Verification:
- scripts/create-oss-release.sh release/oss-2026-06-16
- includes make test-unit: 621 passed, 12 skipped, 26 deselected
- git diff --cached --check
|`SKILLSPECTOR_MODEL`| Override the active provider's bundled default model (see [README.md](../README.md) for per-provider defaults). | `gpt-5.2` |
269
269
270
+
### Live provider tests
271
+
272
+
The manual `test-provider`CI job and local `make test-provider` target perform live requests against provider default endpoints. Missing provider keys print a `WARNING:` line before pytest runs and skip that provider. In CI, missing keys also make the manual job exitwith the configured warning code so GitLab displays the job as passed with warnings; if a key is present but invalid, or the provider request fails, the corresponding test fails.
273
+
274
+
| Command | Required env var | Default URL| Optional model override |
description = "SkillSpector: Security scanner for AI agent skills (Claude Code, Cursor, and similar). Scans skills for vulnerabilities, malicious patterns, and security risks before installation. Supports Git repos, URLs, zips, and local directories; runs static pattern checks and optional LLM semantic analysis; outputs terminal, JSON, and Markdown reports with risk scoring."
9
9
readme = "README.md"
10
10
license = "Apache-2.0"
11
-
requires-python = ">=3.12,<3.14"
11
+
requires-python = ">=3.12,<3.15"
12
12
keywords = [
13
13
"security",
14
14
"ai-agents",
@@ -25,6 +25,7 @@ classifiers = [
25
25
"Programming Language :: Python :: 3",
26
26
"Programming Language :: Python :: 3.12",
27
27
"Programming Language :: Python :: 3.13",
28
+
"Programming Language :: Python :: 3.14",
28
29
"Topic :: Security",
29
30
"Topic :: Software Development :: Quality Assurance",
30
31
]
@@ -38,6 +39,7 @@ dependencies = [
38
39
"openai>=2.25.0",
39
40
"langgraph>=1.0.10",
40
41
"langgraph-cli[inmem]>=0.4.14",
42
+
"langchain-anthropic>=1.4.5",
41
43
"langchain-core>=1.2.17",
42
44
"langchain-openai>=1.1.10",
43
45
"langsmith>=0.7.30",
@@ -97,5 +99,8 @@ source = ["src/skillspector"]
97
99
[tool.pytest.ini_options]
98
100
testpaths = ["tests"]
99
101
asyncio_mode = "auto"
100
-
markers = ["integration: end-to-end tests that invoke the full graph (may call LLMs)"]
101
-
addopts = "-m 'not integration'"
102
+
markers = [
103
+
"integration: end-to-end tests that invoke the full graph (may call LLMs)",
104
+
"provider: live OpenAI/Anthropic/NVIDIA Build provider endpoint tests",
0 commit comments