The Markdown code fence detections for Ruby backtick execution are false positives. The skill does instruct requests to a loopback Obsidian RPC service, which can expose vault metadata and should require user awareness. No prompt injection or remote exfiltration instructions were found.
The skill directs the agent to query a local Obsidian Bases RPC service and shows note paths and frontmatter in results. This can expose private vault data without clear consent.
The skill description and response example show Obsidian base queries returning note paths and metadata. No remote exfiltration is shown, so the risk is local privacy scope.
Capability review items (8)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
The hardcoded 127.0.0.1 address targets a local RPC service that may expose Obsidian data. Loopback scope reduces risk but does not remove the data access concern.
The Markdown code fence detections for Ruby backtick execution are false positives. The skill does instruct requests to a loopback Obsidian RPC service, which can expose vault metadata and should require user awareness. No prompt injection or remote exfiltration instructions were found.
The skill directs the agent to query a local Obsidian Bases RPC service and shows note paths and frontmatter in results. This can expose private vault data without clear consent.
The skill description and response example show Obsidian base queries returning note paths and metadata. No remote exfiltration is shown, so the risk is local privacy scope.
Capability review items (8)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
The hardcoded 127.0.0.1 address targets a local RPC service that may expose Obsidian data. Loopback scope reduces risk but does not remove the data access concern.
Static command and network findings are mostly true positives in documentation examples that call a local Obsidian RPC service with curl and process results with jq. The weak cryptographic algorithm finding is a false positive from the RPC description, and no prompt injection or malicious exfiltration intent was found.
1
Files scanned
65
Lines analyzed
5
Review items
2
False positives ignored
Capability review items (3)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
TRUE POSITIVE: The skill documents curl commands that query a local Obsidian Bases Query RPC endpoint. This is legitimate for the skill purpose, but it can expose structured note metadata and frontmatter to the AI session when a user runs the commands.
The documented commands directly call a local RPC endpoint and return Obsidian base data. The intent is functional rather than malicious, but the data access risk is real.
TRUE POSITIVE: The skill includes jq command examples for extracting fields from RPC responses. The examples are simple and fixed, but agents should not run shell commands against private vault data without user consent.
The shell examples are visible and limited to jq filters. Risk depends on user authorization and the sensitivity of local Obsidian data.
TRUE POSITIVE LOW RISK: The skill links to a GitHub repository for installation through BRAT. This is normal for Obsidian community plugins, but users should review the plugin before installing it.
The GitHub URL is clearly an installation reference, not an automatic download or execution step. Supply-chain risk remains because it depends on an external plugin.
Static false positives ignored (2)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
FALSE POSITIVE FOR EXTERNAL NETWORK RISK: The hardcoded URL and IP address point to 127.0.0.1 for a local Obsidian plugin. No evidence found of outbound third-party exfiltration.
Every detected RPC URL uses 127.0.0.1, which is loopback. The only non-local URL is the installation link for the required plugin.
FALSE POSITIVE: The line describes RPC usage and does not reference a weak hash, cipher, or cryptographic operation. No evidence found of cryptographic code.
Line 3 is plain metadata describing the Bases Query plugin RPC interface. There is no cryptographic algorithm or security primitive in the file.
Static analysis flagged pattern matches for crypto and external commands. Evaluation confirms these are false positives: crypto findings misidentify RPC documentation; command findings are curl examples in docs, not executing code. All network access is to localhost (127.0.0.1:27125) for the Obsidian plugin, a safe local RPC endpoint.
2
Files scanned
442
Lines analyzed
2
Review items
3
False positives ignored
Static false positives ignored (3)
These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.
Static scanner misidentified RPC documentation examples as cryptographic code. The skill contains no cryptographic operations - only local RPC calls to Obsidian plugin.
Static scanner flagged curl examples in documentation as command execution. These are inline documentation examples, not executing code. The skill is a read-only documentation skill with no runtime command execution.
Static scanner flagged localhost URLs for the Obsidian plugin RPC endpoint. These are safe local network references (127.0.0.1:27125) used by the Obsidian Bases Query plugin. No external network access or security risk.
This skill is a pure documentation file (SKILL.md) with no executable code. It describes how to query Obsidian Bases via a local RPC endpoint (127.0.0.1:27125) which only communicates with the local Obsidian application. No network calls to external servers, no file system access beyond its own directory, and no code execution capabilities.
1
Files scanned
65
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.
This skill is a pure documentation file (SKILL.md) with no executable code. It describes how to query Obsidian Bases via a local RPC endpoint (127.0.0.1:27125) which only communicates with the local Obsidian application. No network calls to external servers, no file system access beyond its own directory, and no code execution capabilities.
1
Files scanned
65
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.
This skill is a pure documentation file (SKILL.md) with no executable code. It describes how to query Obsidian Bases via a local RPC endpoint (127.0.0.1:27125) which only communicates with the local Obsidian application. No network calls to external servers, no file system access beyond its own directory, and no code execution capabilities.
1
Files scanned
65
Lines analyzed
0
Review items
0
False positives ignored
No confirmed security findings were recorded for this completed audit.