Back to skill

Security audit

Pipenet-skill

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly a legitimate pipe-network analysis and visualization tool, but it handles user-controlled file paths and generated HTML too loosely for automatic installation without review.

Install only if you are comfortable reviewing and constraining its file access. Use trusted TOML inputs, avoid arbitrary host paths, sanitize network and scenario names, prevent overwrites or path traversal, and open generated HTML only in an isolated browser context until tooltip escaping and output-path validation are fixed.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
src/skill.py:43
Finding

Unrestricted File Access and Path Traversal in Network Operations

Content
View full analysis
Optional[PipeNet]: try: if sys.version_info >= (3, 11): with open(toml_file_path, "rb") as f: data = tomllib.load(f) ``` ### Technical Analysis The public tool interface accepts a caller-controlled `file_path` and uses it directly with `open()` without canonicalization, an approved-root check, or an allowlist. Consequently, the Skill can attempt to read any filesystem path accessible to its operating-system account. The `network.name` value is loaded from attacker-controlled TOML and is subsequently interpolated into output paths. The optional `scenario_name` parameter is likewise interpolated into an HTML output path. Neither value is checked for path separators, `..` components, absolute-path syntax, control characters, or symbolic-link traversal. Although fixed `.toml` and `.html` suffixes limit the names of files that can be written, traversal components can still escape the intended `src/toml` and `src/html` directories. Existing files with compatible suffixes can potentially be overwritten. Symbolic links inside writable output directories may further expand the reachable write scope. Accepting uploaded network files is necessary for the declared functionality, ...[truncated 2179 chars]
Remediation
View remediation
Path: if not user_name or any(c not in "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_-" for c in user_name): raise ValueError("Invalid network name") destination = (INPUT_ROOT / f"{user_name}{suffix}").resolve() if not destination.is_relative_to(INPUT_ROOT): raise ValueError("Path escapes the approved directory") return destination ``` 2. Replace caller-supplied host paths with upload identifiers or server-generated file IDs. Resolve those identifiers through a trusted registry. 3. If path input is unavoidable: - Require the `.toml` extension. - Reject absolute paths, `..`, path separators, NUL bytes, and control characters. - Resolve symbolic links and verify containment after resolution. - Require regular files and reject devices, FIFOs, sockets, and symbolic links. 4. Apply the same validation to `network.name` and `scenario_name`. Prefer server-generated output filenames rather than deriving them from TOML fields. 5. Create output files with exclusive creation where overwriting is unnecessary. Use restrictive permissions and a dedicated, non-executable output directory. 6. Run the Skill under a restricted operating-system account with access only to its upload and output directories. 7. Return generic errors to untrusted callers. Avoid exposing absolute paths, parser details, or unrelated local-file metadata. ]]>

T09 · Insecure Skill Coding Practices

Warning
Location
src/core/visualizer.py:30
Finding

Stored HTML Injection in Generated Network Visualizations

Content
View full analysis
{nid}
类型: 泵站
设定压力: {node.pressure} MPa
计算压力: {attrs['computed_pressure']:.2f} MPa" elif isinstance(node, LoadNode): attrs["demand"] = node.min_flow attrs["title"] = f"{nid}
类型: 负载
需求流量: {node.min_flow}
需求压力: {node.min_pressure}
计算压力: {attrs['computed_pressure']:.2f} MPa" else: attrs["title"] = f"{nid}
类型: 三通
计算压力: {attrs['computed_pressure']:.2f} MPa" G.add_node(nid, **attrs) ``` ### Technical Analysis Node identifiers originate from TOML table keys and are not restricted to a safe identifier syntax. The visualizer inserts `nid` into a `title` property containing HTML markup. Vis-network/PyVis tooltips treat this property as HTML-capable content. JSON serialization may represent angle brackets using Unicode escapes in the generated source, but those escapes are converted back to literal characters when JavaScript evaluates the data. Serialization therefore does not constitute HTML sanitization. An attacker can use a quoted TOML key containing active HTML, such as an image element with an event handler. When the resulting graph is opened and the affected tooltip is rendered, the browser may interpret the injected markup and execute its event handler. This is a stored injection issue because the payload is persisted in the TOML file and then embedded into a generat ...[truncated 1778 chars]
Remediation
View remediation
{safe_nid}
Type: source" ``` 2. Escape all future string-valued TOML attributes used in labels, titles, or generated markup, not only node identifiers. 3. Prefer text-only tooltips. If markup is required, build it from fixed templates and sanitized text nodes rather than concatenating raw strings. 4. Restrict node and edge identifiers to a conservative allowlist such as letters, digits, underscores, periods, and hyphens. Enforce length limits as well. 5. Apply a restrictive Content Security Policy to generated HTML, for example by disallowing inline event handlers and limiting script sources. PyVis generation may need to be adjusted to support nonce- or hash-based scripts. 6. Serve generated visualizations from a sandboxed, untrusted origin. If embedded, use an iframe with a minimal `sandbox` policy and avoid `allow-same-origin` unless strictly necessary. 7. Add regression tests using payloads containing closing tags, image event handlers, SVG event handlers, quotes, and encoded markup. Confirm that each payload is displayed as literal text rather than interpreted as HTML. ]]>
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Memory PoisoningPersistent Context Injection, Context Window Stuffing, Memory Manipulation
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Findings (24)

YARA rule 'agent_skill_mcp_tool_poisoning_metadata': MCP/tool metadata poisoning indicators in tool schemas or skill manifests [agent_skills]

High
Category
YARA Match
Confidence
80% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · src/core/lib/vis-9.1.2/vis-network.min.js (reported line 26)May include surrounding context.

js
),Xl=kl("op-symbols"),Gl=kl("string-to-symbol-registry"),Kl=kl("symbol-to-string-registry"),$l=kl("wks"),Zl=!Rl||!Rl.prototype||!Rl.prototype.findChild,Ql=Kh&&Zh((function(){return 7!=dl(Wl({},"a",{get:function(){return Wl(this,"a",{value:7}).a}})).a}))?function(t,e,i){var n=Hl(Nl,e);n&&delete Nl[e],Wl(t,e,i),n&&t!==Nl&&Wl(Nl,e,n)}:Wl,Jl=function(t,e){var i=Yl[t]=dl(Al);return Bl(i,{type:Dl,tag:t,description:e}),Kh||(i.description=e),i},td=function(t,e,i){t===Nl&&td(Xl,e,i),ol(t);var n=al(e);return ol(i),Qh(Yl,n)?(i.enumerable?(Qh(t,Pl)&&t[Pl][n]&&(t[Pl][n]=!1),i=dl(i,{enumerable:ll(0,!1)})):(Qh(t,Pl)||Wl(t,Pl,ll(1,{})),t[Pl][n]=!0),Ql(t,n,i)):Wl(t,n,i)},ed=function(t,e){ol(t);var i=sl(e),n=cl(i).concat(rd(i));return Ml(n,(function(e){Kh&&!Xh(id,i,e)||td(t,e,i[e])})),t},id=function(t){var e=al(t),i=Xh(Vl,this,e);return!(this===Nl&&Qh(Yl,e)&&!Qh(Xl,e))&&(!(i||!Qh(this,e)||!Qh(Yl,e)||Qh(this,Pl)&&this[Pl][e])||i)},nd=function(t,e){var i=sl(t),n=al(e);if(i!==Nl||!Qh(Yl,n)||Qh(Xl,n)){var o

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

All user-facing instructions and trigger descriptions are written only in Chinese, with no indication that other languages are supported or that the Chinese-only constraint is intentional and justified. This can violate language or locale policy when a skill implicitly mandates a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill description says it will convert model output into a TOML file, which affects user data or the local filesystem, but no warning or disclosure is provided about file creation or overwrite behavior. For markdown skill descriptions, file-affecting behavior should be disclosed so users understand the impact before using the skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The usage section says only '告诉用户如何触发该技能' and then lists example commands, but it does not clarify whether these are the only valid triggers, whether variations are allowed, or when the skill should not activate. This ambiguity can cause unintended invocation because the trigger scope and constraints are not explicitly bounded.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The manifest advertises generation of a TOML configuration file but does not warn that this may create or overwrite files, nor does it describe safe handling of output paths. In practice, users or agents may assume the action is purely in-memory, leading to unintended persistence, overwriting existing configurations, or unsafe chaining into file-path-based tools.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The tool description says to call the network-design function whenever the user wants to create or modify a network, but it does not define clear guardrails such as requiring confirmation before writing files or restricting when side-effecting actions are allowed. In an agent setting, broad invocation criteria can cause the model to generate or overwrite configuration artifacts based on ambiguous prompts, increasing the risk of unintended file operations or unsafe downstream analysis.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

This Python file contains Chinese-only docstrings, comments, and report strings such as the analyzer title and printed report labels. Under the policy, forcing a specific language without opt-in is a natural-language policy violation unless the locale restriction is explicitly justified, which is not present here.

Content

No source excerpt is available for this finding.

Context Window Stuffing

Medium
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.

Content

Scanner excerpt · src/core/lib/vis-9.1.2/vis-network.css (reported line 1)May include surrounding context.

css
.vis-overlay{bottom:0;left:0;position:absolute;right:0;top:0;z-index:10}.vis-active{box-shadow:0 0 10px #86d5f8}.vis [class*=span]{min-height:0;width:auto}div.vis-color-picker{background-color:#fff;border-radius:15px;box-shadow:0 0 10px 0 rgba(0,0,0,.5);display:none;height:444px;left:30px;margin-left:30px;margin-top:-140px;padding:10px;position:absolute;top:0;width:310px;z-index:1}div.vis-color-picker div.vis-arrow{left:5px;position:absolute;top:147px}div.vis-color-picker div.vis-arrow:after,div.vis-color-picker div.vis-arrow:before{border:solid transparent;content:" ";height:0;pointer-events:none;position:absolute;right:100%;top:50%;width:0}div.vis-color-picker div.vis-arrow:after{border-color:hsla(0,0%,100%,0) #fff hsla(0,0%,100%,0) hsla(0,0%,100%,0);border-width:30px;margin-top:-30px}div.vis-color-picker div.vis-color{cursor:pointer;height:289px;position:absolute;width:289px}div.vis-color-picker div.vis-brightness{position:absolute;top:313px}div.vis-color-picker div.vis-opacity{position:absolute;top:350px}div.vis-color-picker div.vis-selector{background:#4c4c4c;background:-moz-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:-webkit-gradient(linear,left top,left bottom,color-stop(0,#4c4c4c),color-stop(12%,#595959),color-stop(25%,#666),color-stop(39%,#474747),color-stop(50%,#2c2c2c),color-stop(51%,#000),color-stop(60%,#111),color-stop(76%,#2b2b2b),color-stop(91%,#1c1c1c),color-stop(100%,#131313));background:-webkit-linear-gradient(top,#4c4c4c,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313);background:-o-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:-ms-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:linear-gradient(180deg,#4c4c4c 0,#595959
...[truncated 28 chars]

Context Window Stuffing

Medium
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.

Content

Scanner excerpt · src/core/lib/vis-9.1.2/vis-network.css (reported line 1)May include surrounding context.

css
.vis-overlay{bottom:0;left:0;position:absolute;right:0;top:0;z-index:10}.vis-active{box-shadow:0 0 10px #86d5f8}.vis [class*=span]{min-height:0;width:auto}div.vis-color-picker{background-color:#fff;border-radius:15px;box-shadow:0 0 10px 0 rgba(0,0,0,.5);display:none;height:444px;left:30px;margin-left:30px;margin-top:-140px;padding:10px;position:absolute;top:0;width:310px;z-index:1}div.vis-color-picker div.vis-arrow{left:5px;position:absolute;top:147px}div.vis-color-picker div.vis-arrow:after,div.vis-color-picker div.vis-arrow:before{border:solid transparent;content:" ";height:0;pointer-events:none;position:absolute;right:100%;top:50%;width:0}div.vis-color-picker div.vis-arrow:after{border-color:hsla(0,0%,100%,0) #fff hsla(0,0%,100%,0) hsla(0,0%,100%,0);border-width:30px;margin-top:-30px}div.vis-color-picker div.vis-color{cursor:pointer;height:289px;position:absolute;width:289px}div.vis-color-picker div.vis-brightness{position:absolute;top:313px}div.vis-color-picker div.vis-opacity{position:absolute;top:350px}div.vis-color-picker div.vis-selector{background:#4c4c4c;background:-moz-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:-webkit-gradient(linear,left top,left bottom,color-stop(0,#4c4c4c),color-stop(12%,#595959),color-stop(25%,#666),color-stop(39%,#474747),color-stop(50%,#2c2c2c),color-stop(51%,#000),color-stop(60%,#111),color-stop(76%,#2b2b2b),color-stop(91%,#1c1c1c),color-stop(100%,#131313));background:-webkit-linear-gradient(top,#4c4c4c,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313);background:-o-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:-ms-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:linear-gradient(180deg,#4c4c4c 0,#595959
...[truncated 28 chars]

Context Window Stuffing

Medium
Category
Memory Poisoning
Confidence
80% confidence
Finding

Skill attempts to fill the context window with filler content, displacing legitimate instructions and safety constraints. This can degrade agent performance or bypass safety boundaries.

Content

Scanner excerpt · src/core/lib/vis-9.1.2/vis-network.css (reported line 1)May include surrounding context.

css
.vis-overlay{bottom:0;left:0;position:absolute;right:0;top:0;z-index:10}.vis-active{box-shadow:0 0 10px #86d5f8}.vis [class*=span]{min-height:0;width:auto}div.vis-color-picker{background-color:#fff;border-radius:15px;box-shadow:0 0 10px 0 rgba(0,0,0,.5);display:none;height:444px;left:30px;margin-left:30px;margin-top:-140px;padding:10px;position:absolute;top:0;width:310px;z-index:1}div.vis-color-picker div.vis-arrow{left:5px;position:absolute;top:147px}div.vis-color-picker div.vis-arrow:after,div.vis-color-picker div.vis-arrow:before{border:solid transparent;content:" ";height:0;pointer-events:none;position:absolute;right:100%;top:50%;width:0}div.vis-color-picker div.vis-arrow:after{border-color:hsla(0,0%,100%,0) #fff hsla(0,0%,100%,0) hsla(0,0%,100%,0);border-width:30px;margin-top:-30px}div.vis-color-picker div.vis-color{cursor:pointer;height:289px;position:absolute;width:289px}div.vis-color-picker div.vis-brightness{position:absolute;top:313px}div.vis-color-picker div.vis-opacity{position:absolute;top:350px}div.vis-color-picker div.vis-selector{background:#4c4c4c;background:-moz-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:-webkit-gradient(linear,left top,left bottom,color-stop(0,#4c4c4c),color-stop(12%,#595959),color-stop(25%,#666),color-stop(39%,#474747),color-stop(50%,#2c2c2c),color-stop(51%,#000),color-stop(60%,#111),color-stop(76%,#2b2b2b),color-stop(91%,#1c1c1c),color-stop(100%,#131313));background:-webkit-linear-gradient(top,#4c4c4c,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313);background:-o-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:-ms-linear-gradient(top,#4c4c4c 0,#595959 12%,#666 25%,#474747 39%,#2c2c2c 50%,#000 51%,#111 60%,#2b2b2b 76%,#1c1c1c 91%,#131313 100%);background:linear-gradient(180deg,#4c4c4c 0,#595959
...[truncated 28 chars]

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · src/core/lib/vis-9.1.2/vis-network.min.js (reported line 26)May include surrounding context.

js
*
 * vis.js may be distributed under either license.
 */
!function(t,e){"object"==typeof exports&&"undefined"!=typeof module?e(exports):"function"==typeof define&&define.amd?define(["exports"],e):e((t="undefined"!=typeof globalThis?globalThis:t||self).vis=t.vis||{})}(this,(function(t){"use strict";var e="undefined"!=typeof globalThis?globalThis:"undefined"!=typeof window?window:"undefined"!=typeof global?global:"undefined"!=typeof self?self:{},i=function(t){return t&&t.Math==Math&&t},n=i("object"==typeof globalThis&&globalThis)||i("object"==typeof window&&window)||i("object"==typeof self&&self)||i("object"==typeof e&&e)||function(){return this}()||Function("return this")(),o=function(t){try{return!!t()}catch(t){return!0}},r=!o((function(){var t=function(){}.bind();return"function"!=typeof t||t.hasOwnProperty("prototype")})),s=r,a=Function.prototype,h=a.apply,l=a.call,d="object"==typeof Reflect&&Reflect.apply||(s?l.bind(h):function(){return l.apply(h,arguments)}),c=r,u=Function.prototype,f=u.bind,p=u.call,v=c&&f.bind(p,p),g=c?function(t){return t&&v(t)}:function(t){return t&&function(){return p.apply(t,arguments)}},y=function(t){return"function"==typeof t},m={},b=!o((function(){return 7!=Object.defineProperty({},1,{get:function(){return 7}})[1]})),w=r,k=Function.prototype.call,_=w?k.bind(k):function(){return k.apply(k,arguments)},x={},E={}.propertyIsEnumerable,O=Object.getOwnPropertyDescriptor,C=O&&!E.call({1:2},1);x.f=C?function(t){var e=O(this,t);return!!e&&e.enumerable}:E;var S,T,M=function(t,e){return{enumerable:!(1&t),configurable:!(2&t),writable:!(4&t),value:e}},P=g,D=P({}.toString),I=P("".slice),B=function(t){return I(D(t),8,-1)},z=g,N=o,F=B,A=n.Object,j=z("".split),R=N((function(){return!A("z").propertyIsEnumerable(0)}))?function(t){return"String"==F(t)?j(t,""):A(t)}:A,L=n.TypeError,H=function(t){if(null==t)throw L("Can't call method on "+t);return t},W=R,q=H,V=function(t){return W(q(t))},U=y,Y=function(t){return"object"==typeof t?null!==t:U(t)},X={},G=X,K=n,
...[truncated 28 chars]

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · src/core/lib/vis-9.1.2/vis-network.min.js (reported line 28)May include surrounding context.

js
*/
!function(t,e){"object"==typeof exports&&"undefined"!=typeof module?e(exports):"function"==typeof define&&define.amd?define(["exports"],e):e((t="undefined"!=typeof globalThis?globalThis:t||self).vis=t.vis||{})}(this,(function(t){"use strict";var e="undefined"!=typeof globalThis?globalThis:"undefined"!=typeof window?window:"undefined"!=typeof global?global:"undefined"!=typeof self?self:{},i=function(t){return t&&t.Math==Math&&t},n=i("object"==typeof globalThis&&globalThis)||i("object"==typeof window&&window)||i("object"==typeof self&&self)||i("object"==typeof e&&e)||function(){return this}()||Function("return this")(),o=function(t){try{return!!t()}catch(t){return!0}},r=!o((function(){var t=function(){}.bind();return"function"!=typeof t||t.hasOwnProperty("prototype")})),s=r,a=Function.prototype,h=a.apply,l=a.call,d="object"==typeof Reflect&&Reflect.apply||(s?l.bind(h):function(){return l.apply(h,arguments)}),c=r,u=Function.prototype,f=u.bind,p=u.call,v=c&&f.bind(p,p),g=c?function(t){return t&&v(t)}:function(t){return t&&function(){return p.apply(t,arguments)}},y=function(t){return"function"==typeof t},m={},b=!o((function(){return 7!=Object.defineProperty({},1,{get:function(){return 7}})[1]})),w=r,k=Function.prototype.call,_=w?k.bind(k):function(){return k.apply(k,arguments)},x={},E={}.propertyIsEnumerable,O=Object.getOwnPropertyDescriptor,C=O&&!E.call({1:2},1);x.f=C?function(t){var e=O(this,t);return!!e&&e.enumerable}:E;var S,T,M=function(t,e){return{enumerable:!(1&t),configurable:!(2&t),writable:!(4&t),value:e}},P=g,D=P({}.toString),I=P("".slice),B=function(t){return I(D(t),8,-1)},z=g,N=o,F=B,A=n.Object,j=z("".split),R=N((function(){return!A("z").propertyIsEnumerable(0)}))?function(t){return"String"==F(t)?j(t,""):A(t)}:A,L=n.TypeError,H=function(t){if(null==t)throw L("Can't call method on "+t);return t},W=R,q=H,V=function(t){return W(q(t))},U=y,Y=function(t){return"object"==typeof t?null!==t:U(t)},X={},G=X,K=n,$=y,Z=function(t){return $(t)?t:void 0},Q=function(t,e
...[truncated 28 chars]

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

Node and edge titles shown to users contain Chinese strings such as labels for type, pressure, flow, and status. Because the file presents a fixed language in user-visible content with no opt-in, toggle, or justification, it violates the language/locale policy criterion.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The class docstring and user-facing messages are written in Chinese throughout the skill, including entry-point descriptions and status/error text. There is no indication that users may choose another language or that the Chinese-only locale is a documented requirement for a region-specific tool.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The documentation states this entrypoint 'receives an Agent-generated TOML string' and that the tool 'generates a TOML configuration string'. In reality, the implementation parses the provided TOML, loads it into a pipe network object, and persists the content to ./src/toml/{pipe_net.name}.toml, which is a materially different behavior with file-system side effects.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The code writes attacker-controlled content to a path derived from pipe_net.name without sanitization or confinement checks. If pipe_net.name can contain path traversal sequences or special filenames, an attacker may overwrite arbitrary files accessible to the process, making the file-write side effect materially dangerous in this skill context.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The visualization output path is built from pipe_net.name and scenario_name, both influenced by input, and then written to disk as HTML. Without sanitization and path validation, this can enable arbitrary file creation/overwrite via traversal or crafted filenames; additionally, generated HTML may become an XSS vector if later served in a browser and embeds unsanitized data.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

This file’s natural-language interface content, including the main module description and usage guidance, is entirely in Chinese. Under the policy, locale-specific language is only acceptable when the skill offers user choice or clearly documents a justified region-specific constraint, which is not present here.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The documented return structure lists a 'flow_paths' field as a key output of the comprehensive analysis. However, the actual return value at L45-L50 does not include 'flow_paths' and instead exposes 'other_nodes' and 'all_edges', which contradicts the stated interface and intent of the method documentation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

This code contains multiple natural-language docstrings and print messages in Chinese, such as the class docstring and success/error output. That can violate a language/locale policy when the skill forces a specific language without user opt-in or documented justification.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
95% confidence
Finding

The class docstring and all user-visible log strings are written only in Chinese, which imposes a specific language on users without offering a choice or documenting a region-specific justification. This matches the natural-language policy violation category for language or locale constraints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

This code emits a warning string in Chinese directly to users, which enforces a specific language in a general-purpose skill file. The same pattern appears elsewhere in the file, and there is no opt-in, fallback, or documentation indicating the skill is intentionally limited to Chinese-speaking users.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
77% confidence
Finding

With no manifest available, the skill's intended scope is unknown. This file is framed as a visualizer and its method generates and saves an interactive HTML file, which is a filesystem-writing capability not otherwise justified by any stated purpose or permissions context.

Content

No source excerpt is available for this finding.