Most static alerts are false positives caused by Markdown fences, inline code, emoji, and fixed relative documentation links. Two instructions intentionally run bundled Python connectors, and one accesses a public YouTube feed; these are expected features but create limited command and network exposure.
2
Files scanned
551
Lines analyzed
7
Review items
0
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.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
The line explicitly instructs the agent to run a Python GDELT connector. The competitor placeholder is quoted, but executing an unreviewed bundled script still crosses a command-execution boundary.
The line explicitly runs a Python YouTube connector and describes piping RSS data into another script. These intended operations execute code not included in the reviewed files.
The workflow directs the agent to retrieve a public YouTube RSS endpoint or use the YouTube connector. This expected feature discloses network metadata and requested channel identifiers externally.
Most static alerts are false positives caused by Markdown fences, inline code, emoji, and fixed relative documentation links. Two instructions intentionally run bundled Python connectors, and one accesses a public YouTube feed; these are expected features but create limited command and network exposure.
2
Files scanned
551
Lines analyzed
7
Review items
0
False positives ignored
Capability review items (3)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
The line explicitly instructs the agent to run a Python GDELT connector. The competitor placeholder is quoted, but executing an unreviewed bundled script still crosses a command-execution boundary.
The line explicitly runs a Python YouTube connector and describes piping RSS data into another script. These intended operations execute code not included in the reviewed files.
The workflow directs the agent to retrieve a public YouTube RSS endpoint or use the YouTube connector. This expected feature discloses network metadata and requested channel identifiers externally.
Most static findings are false positives caused by Markdown fences, relative documentation links, connector labels, and metadata URLs. Two command recipes are confirmed because they execute local Python connectors with user-supplied placeholders; the channel handle is unquoted and can permit shell injection. No prompt injection or data-exfiltration intent was found.
2
Files scanned
551
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (2)
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 skill directs execution of a Python connector with an unquoted user-provided partner handle. Shell metacharacters in that value could inject additional commands.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
The skill directs execution of a local Python connector and interpolates a user-provided competitor query into shell text. Crafted query content could alter shell parsing if substituted without strict argument handling.
Most scanner hits are false positives from Markdown code fences, relative documentation links, and table text. Two hits are confirmed because the skill documents optional Python connector commands, and one hit confirms external YouTube RSS/API network access. No prompt injection or malicious data exfiltration intent was found in the reviewed files.
2
Files scanned
551
Lines analyzed
7
Review items
0
False positives ignored
Capability review items (3)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
Line 59 documents a python3 command that runs a local GDELT connector script with a competitor argument. It is legitimate tooling, but it is still an external command with network-facing behavior.
Line 61 documents a python3 command for a YouTube connector and references RSS monitoring. This is a real external command path, even though it is scoped to campaign research.
Most static findings are markdown links, prompt examples, connector labels, or project metadata. I confirmed two explicit Python connector command templates as medium external-command risks because they run local scripts with user-supplied competitor or partner values. No prompt injection or malicious exfiltration intent was found.
2
Files scanned
551
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (2)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
This line gives a concrete python3 command template that runs a local connector script with a user-supplied competitor value. It is legitimate tooling, but external command execution should require confirmation and input sanitization.
This line gives a concrete python3 command template for a YouTube connector using a partner handle. It is a real external command path and should be guarded against unsafe argument handling.
Most static findings are markdown links, prompt examples, connector labels, or project metadata. I confirmed two explicit Python connector command templates as medium external-command risks because they run local scripts with user-supplied competitor or partner values. No prompt injection or malicious exfiltration intent was found.
2
Files scanned
551
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (2)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
This line gives a concrete python3 command template that runs a local connector script with a user-supplied competitor value. It is legitimate tooling, but external command execution should require confirmation and input sanitization.
This line gives a concrete python3 command template for a YouTube connector using a partner handle. It is a real external command path and should be guarded against unsafe argument handling.
No prompt-injection or data-exfiltration intent was found in SKILL.md or references/templates.md. Most static findings are Markdown links, code fences, connector placeholders, or public documentation URLs. The only confirmed risks are optional shell commands in SKILL.md lines 59 and 61, which should require user confirmation before execution.
2
Files scanned
551
Lines analyzed
6
Review items
0
False positives ignored
Capability review items (2)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
**Keyless news read on rivals**: `python3 "${CLAUDE_PLUGIN_ROOT}/scripts/connectors/gdelt.py" '"<com
Line 59 instructs running python3 with a local gdelt.py connector and a user-supplied competitor string. It is an intentional external command with network-facing behavior, so host execution should require confirmation.
Line 61 instructs running python3 with a YouTube connector and mentions piping a YouTube RSS feed into rss_monitor.py. These are explicit external commands and data fetches, so they should not auto-run without consent.