Security audit

AI Tech Intelligence Briefing

Security checks for vulnerabilities and agentic risk

Overview

This briefing skill is not malware, but it needs Review because it can read unexpected Markdown files and its documentation overstates what the code actually does.

Review this package before installing in shared or sensitive workspaces. Treat generated briefings as demo content, expect an embedded donation/footer unless modified, and avoid passing untrusted values to the fetch command until date validation and path containment are added.

A.I.G

Vulnerability Patterns

Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review

Findings (3)

other

Error

Location: scripts/briefing.py:83
Finding: Forced Promotional Content in Every Generated Briefing
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: scripts/briefing.py:140
Finding: Path Traversal Allows Reading Markdown Files Outside the Briefings Directory
Content: View full analysis
str: """Fetch saved briefing.""" filename = self.briefings_dir / f"{date}.md" if filename.exists(): return filename.read_text(encoding='utf-8') return None ``` The unvalidated argument is passed directly to the vulnerable method: ```python elif command == 'fetch': if len(args) < 2: print("❌ Error: Please specify date (YYYY-MM-DD)") sys.exit(1) briefing = generator.fetch_briefing(args[1]) if briefing: print(briefing) ``` ### Technical Analysis Although the command documentation requires a date in `YYYY-MM-DD` format, the implementation does not validate the argument. `pathlib.Path` preserves traversal components such as `../`, allowing the constructed path to escape `self.briefings_dir`. The forced `.md` suffix limits exploitation to filenames ending in `.md`, but does not prevent access to Markdown files elsewhere in the current workspace or filesystem. The code only checks whether the resulting path exists; it does not resolve the path and verify that it remains under the intended briefing directory. ### Attack Path 1. An attacker influences the argument supplied to the `fetch` command. 2. The attacker supplies traversal input such as: ```text fetch ../../private/notes ``` 3. The code constructs: ```text /briefings/../../private/notes.md ``` 4. The operating system normalizes the traversal components. 5. If the target exists and is readable by the process, its contents are loaded. 6. The command prints the contents to standard output, potentially exposing them to an Agent transcript or downstream user. ### Impact Assessment An attacker can read arbitrary accessible files whose names end ...[truncated 402 chars]
Remediation: View remediation

T08 · Insecure Dependencies

Note

Location: setup.py:28
Finding: Unused and Unlocked HTTP Dependency Expands the Supply-Chain Attack Surface
Content: View full analysis
=2.28.0", ], ``` The package manifest also declares the dependency using nonstandard npm dependency notation: ```json "dependencies": { "requests>=2.31.0": "Python package for HTTP requests" }, ``` ### Technical Analysis The executable does not import or use `requests`; it generates hard-coded demo stories and performs no network retrieval. Installing the library therefore provides no functionality while introducing additional third-party and transitive code into the environment. The Python requirement specifies only a lower bound, so installations can resolve to different future releases rather than a reviewed, reproducible dependency set. The `package.json` declaration is also malformed for an npm dependency map: the package name contains a Python-style version constraint, while the value is a description rather than a package version. No evidence shows that the currently referenced `requests` package is malicious. The risk arises from unnecessary dependency installation, inconsistent package-manager metadata, and the absence of a reviewed lock file or hashes. ### Attack Path 1. A user installs the project through its Python packaging metadata. 2. The resolver downloads `requests` and its transitive dependencies despite the application not using them. 3. Dependency versions are selected according to the state of the package index at installation time. 4. If a future resolved package or transitive component is compromised, malicious installation or runtime behavior could execute with the installer's privileges. 5. The malformed `package.json` declaration may additionally cause installation failure or unexpected package-manager behavior. ### Impact Assessment The issue unnecessarily ...[truncated 293 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The trigger phrase "daily briefing" is broad and generic, making it likely to match ordinary user requests that are not specifically intended to invoke this skill. This can cause unintended activation and execution of network-fetching behavior, which increases the risk of user confusion, overreach, and accidental data handling beyond user expectations.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The top-level docstring states that the script generates daily AI/tech briefings using public RSS feeds and APIs. In practice, generate_briefing calls _generate_demo_stories, and the code explicitly notes it is generating demo content instead of fetching real data, so the documentation materially misrepresents the tool's actual behavior.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 76% confidence
Finding: The comment '# Get name' suggests the code is retrieving the package name, but the jq expression reads '.description | split(":")[0]' from package.json. This is an intent/documentation mismatch because the inline comment states one behavior while the code performs a different metadata extraction.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.