📦
Historial de auditorías
qryma-search - 7 auditorías
Comparación de versiones
Cambios de capacidades y hallazgos entre versiones auditadas, primero las más recientes.
| Versión | Fecha | Resultado | Elementos de revisión | Cambio vs anterior |
|---|---|---|---|---|
| v7 Más reciente | 5 jul 2026, 21:16 | 9 confirmado | 35 | Sin cambios de capacidad |
| v6 | 5 jul 2026, 21:16 | 9 confirmado | 35 | Sin cambios de capacidad |
| v5 | 30 jun 2026, 11:29 | 1 confirmado | 4 | Comandos externosAcceso al sistema de archivos |
| v4 | 2 abr 2026, 02:51 | 1 confirmado | 0 | Acceso al sistema de archivos |
| v3 | 26 mar 2026, 02:31 | Sin hallazgos confirmados | 1 | Sin cambios de capacidad |
| v2 | 25 mar 2026, 09:16 | Sin hallazgos confirmados | 2 | Sin cambios de capacidad |
| v1 | 25 mar 2026, 08:14 | Sin hallazgos confirmados | 4 | Base |
Versión de auditoría 7 Más reciente
5 jul 2026, 21:16
The skill is a legitimate Qryma web-search wrapper, but it performs real network egress and handles API credentials. Confirmed risks include API key loading from environment and .env files, transmission of the key in an outbound header, a configurable endpoint that can redirect credentials, and documentation that encourages sharing keys with an AI assistant.
12
Archivos escaneados
770
Líneas analizadas
48
Elementos de revisión
0
Falsos positivos ignorados
Preocupaciones de seguridad confirmadas (9)
Alto
Environment file access
# Method 2: Create .env file
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Environment file access
echo 'QRYMA_API_KEY=ak-your-api-key-here' > .env
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Environment file access
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
env_path = ".env"
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
env_path = ".env"
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
Create `.env` file:
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
API Key Sharing Guidance Risks Secret Disclosure
README.md and SKILL.md tell users to ask an AI assistant to configure QRYMA_API_KEY. If users replace the placeholder with a real key, the secret can be stored in chat logs, tool transcripts, or third-party systems.
The documentation explicitly connects QRYMA_API_KEY setup with asking an AI assistant to configure it. This is a clear social-engineering risk for credential exposure.
Alto
Configurable Endpoint Can Receive API Key
scripts/search_core.py loads QRYMA_ENDPOINT from the environment or .env files, then sends self.api_key in the X-Api-Key header to that endpoint. A modified endpoint can receive both user queries and the API key.
The data flow is visible in the source: endpoint configuration is loaded before the request is built, and the API key is attached to the request headers.
Elementos de revisión de capacidades (35)
Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.
Alto
Generic API/secret keys · 2 apariciones
adapter = QrymaAdapter(api_key=getattr(args, "api_key", None))
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
api_key = event.get("api_key") if isinstance(event, dict) else None
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
adapter = QrymaAdapter(api_key=api_key)
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys · 2 apariciones
Get your free QRYMA_API_KEY from [qryma.com](https://qryma.com/). Chat with your AI assistant and as
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys · 2 apariciones
Please configure qryma search with the QRYMA_API_KEY set to ak-your-api-key-here.
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
从[qryma.com](https://qryma.com/) 获取免费的QRYMA_API_KEY,直接发送给你的AI助手,让他帮你自动配置。
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
请帮我配置好qryma search,设置QRYMA_API_KEY为ak-your-api-key-here
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys · 2 apariciones
export QRYMA_API_KEY="ak-your-api-key-here"
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Generic API/secret keys
echo 'QRYMA_API_KEY=ak-your-api-key-here' > .env
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Generic API/secret keys
def __init__(self, core: QrymaSearchCore = None, api_key: str = None):
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
self.core = QrymaSearchCore(api_key=api_key)
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Hidden file in home directory · 2 apariciones
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Hidden file access · 2 apariciones
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Python environment access
key = os.environ.get("QRYMA_API_KEY")
The code reads QRYMA_API_KEY from the process environment for use as a credential. This is expected for the skill, but it is still real environment-secret access.
Alto
Generic API/secret keys
key = os.environ.get("QRYMA_API_KEY")
The code reads QRYMA_API_KEY from the process environment for use as a credential. This is expected for the skill, but it is still real environment-secret access.
Alto
Generic API/secret keys
m = re.search(r"^\s*QRYMA_API_KEY\s*=\s*(.+?)\s*$", txt, re.M)
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Generic API/secret keys
def __init__(self, api_key: Optional[str] = None, endpoint: Optional[str] = None):
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
self.api_key = api_key or load_key()
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
"X-Api-Key": self.api_key,
The code places self.api_key into the X-Api-Key header on the outbound request. This transmits a secret externally and becomes an exfiltration path if the endpoint is redirected.
Alto
Generic API/secret keys
从 [qryma.com](https://qryma.com/) 获取免费的 QRYMA_API_KEY,直接发送给你的 AI 助手,让他帮你自动配置。
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
请帮我配置好 qryma search,设置 QRYMA_API_KEY 为 ak-your-api-key-here
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
QRYMA_API_KEY=ak-your-api-key-here
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Medio
Python environment access
endpoint = os.environ.get("QRYMA_ENDPOINT")
QRYMA_ENDPOINT from the environment controls the destination used for the later request. If this setting is changed unexpectedly, the API key and queries can be sent to an untrusted service.
Bajo
Hardcoded URL
"default": "https://search.qryma.com/api/web"
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Python HTTP libraries
import urllib.request
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Python HTTP libraries
req = urllib.request.Request(
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Python HTTP libraries
with urllib.request.urlopen(req, timeout=30) as resp:
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Hardcoded URL
QRYMA_URL = "https://search.qryma.com/api/web"
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Hardcoded URL
return "https://search.qryma.com/api/web"
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Factores de riesgo
🔑 Variables de entorno (32)
main_claw.py:35 main_claw.py:41 main_claw.py:42 manifest.json:99 README.md:21 README.md:24 README.md:27 README.md:30 README.md:37 README.md:40 requirements.txt:2 scripts/adapters/adapter.py:28 scripts/adapters/adapter.py:73 scripts/adapters/adapter.py:77 scripts/search_core.py:19 scripts/search_core.py:24 scripts/search_core.py:50 scripts/search_core.py:19 scripts/search_core.py:36 scripts/search_core.py:77 scripts/search_core.py:78 scripts/search_core.py:91 scripts/search_core.py:92 scripts/search_core.py:119 skill.json:6 skill.json:7 SKILL.md:46 SKILL.md:49 SKILL.md:52 SKILL.md:55 SKILL.md:62 SKILL.md:68
🌐 Acceso a red (17)
manifest.json:7 manifest.json:57 manifest.json:107 package.json:22 package.json:24 README.md:12 README.md:15 scripts/search_core.py:9 scripts/search_core.py:112 scripts/search_core.py:123 scripts/search_core.py:14 scripts/search_core.py:71 scripts/search_core.py:118 SKILL.md:12 SKILL.md:38 SKILL.md:46 SKILL.md:52
📁 Acceso al sistema de archivos (4)
Auditado por: codex
5 jul 2026, 21:16
The skill is a legitimate Qryma web-search wrapper, but it performs real network egress and handles API credentials. Confirmed risks include API key loading from environment and .env files, transmission of the key in an outbound header, a configurable endpoint that can redirect credentials, and documentation that encourages sharing keys with an AI assistant.
12
Archivos escaneados
770
Líneas analizadas
48
Elementos de revisión
0
Falsos positivos ignorados
Preocupaciones de seguridad confirmadas (9)
Alto
Environment file access
# Method 2: Create .env file
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Environment file access
echo 'QRYMA_API_KEY=ak-your-api-key-here' > .env
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Environment file access
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
env_path = ".env"
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
env_path = ".env"
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Environment file access
Create `.env` file:
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
API Key Sharing Guidance Risks Secret Disclosure
README.md and SKILL.md tell users to ask an AI assistant to configure QRYMA_API_KEY. If users replace the placeholder with a real key, the secret can be stored in chat logs, tool transcripts, or third-party systems.
The documentation explicitly connects QRYMA_API_KEY setup with asking an AI assistant to configure it. This is a clear social-engineering risk for credential exposure.
Alto
Configurable Endpoint Can Receive API Key
scripts/search_core.py loads QRYMA_ENDPOINT from the environment or .env files, then sends self.api_key in the X-Api-Key header to that endpoint. A modified endpoint can receive both user queries and the API key.
The data flow is visible in the source: endpoint configuration is loaded before the request is built, and the API key is attached to the request headers.
Elementos de revisión de capacidades (35)
Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.
Alto
Generic API/secret keys · 2 apariciones
adapter = QrymaAdapter(api_key=getattr(args, "api_key", None))
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
api_key = event.get("api_key") if isinstance(event, dict) else None
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
adapter = QrymaAdapter(api_key=api_key)
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys · 2 apariciones
Get your free QRYMA_API_KEY from [qryma.com](https://qryma.com/). Chat with your AI assistant and as
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys · 2 apariciones
Please configure qryma search with the QRYMA_API_KEY set to ak-your-api-key-here.
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
从[qryma.com](https://qryma.com/) 获取免费的QRYMA_API_KEY,直接发送给你的AI助手,让他帮你自动配置。
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
请帮我配置好qryma search,设置QRYMA_API_KEY为ak-your-api-key-here
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys · 2 apariciones
export QRYMA_API_KEY="ak-your-api-key-here"
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Generic API/secret keys
echo 'QRYMA_API_KEY=ak-your-api-key-here' > .env
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Alto
Generic API/secret keys
def __init__(self, core: QrymaSearchCore = None, api_key: str = None):
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
self.core = QrymaSearchCore(api_key=api_key)
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Hidden file in home directory · 2 apariciones
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Hidden file access · 2 apariciones
env_path = os.path.expanduser("~/.qryma/.env")
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Python environment access
key = os.environ.get("QRYMA_API_KEY")
The code reads QRYMA_API_KEY from the process environment for use as a credential. This is expected for the skill, but it is still real environment-secret access.
Alto
Generic API/secret keys
key = os.environ.get("QRYMA_API_KEY")
The code reads QRYMA_API_KEY from the process environment for use as a credential. This is expected for the skill, but it is still real environment-secret access.
Alto
Generic API/secret keys
m = re.search(r"^\s*QRYMA_API_KEY\s*=\s*(.+?)\s*$", txt, re.M)
The code reads .env or ~/.qryma/.env and extracts QRYMA credentials or endpoint settings. Automatic reads from hidden or local environment files can expose secrets outside the skill package boundary.
Alto
Generic API/secret keys
def __init__(self, api_key: Optional[str] = None, endpoint: Optional[str] = None):
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
self.api_key = api_key or load_key()
The code accepts an API key through CLI or event parameters and passes it into the adapter/core. Secrets provided this way can appear in process listings, shell history, captured events, or logs.
Alto
Generic API/secret keys
"X-Api-Key": self.api_key,
The code places self.api_key into the X-Api-Key header on the outbound request. This transmits a secret externally and becomes an exfiltration path if the endpoint is redirected.
Alto
Generic API/secret keys
从 [qryma.com](https://qryma.com/) 获取免费的 QRYMA_API_KEY,直接发送给你的 AI 助手,让他帮你自动配置。
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
请帮我配置好 qryma search,设置 QRYMA_API_KEY 为 ak-your-api-key-here
The documentation tells users to provide QRYMA_API_KEY through an AI assistant prompt or similar chat flow. That can disclose a real API key into chat logs, tool transcripts, or third-party model systems.
Alto
Generic API/secret keys
QRYMA_API_KEY=ak-your-api-key-here
The documentation shows storing QRYMA_API_KEY in shell environment state or a plaintext .env file. This is a common setup method, but it creates a real local secret-exposure risk if history or files are not protected.
Medio
Python environment access
endpoint = os.environ.get("QRYMA_ENDPOINT")
QRYMA_ENDPOINT from the environment controls the destination used for the later request. If this setting is changed unexpectedly, the API key and queries can be sent to an untrusted service.
Bajo
Hardcoded URL
"default": "https://search.qryma.com/api/web"
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Python HTTP libraries
import urllib.request
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Python HTTP libraries
req = urllib.request.Request(
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Python HTTP libraries
with urllib.request.urlopen(req, timeout=30) as resp:
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Hardcoded URL
QRYMA_URL = "https://search.qryma.com/api/web"
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Bajo
Hardcoded URL
return "https://search.qryma.com/api/web"
The skill performs outbound HTTP requests to the Qryma search API and sends user search queries off-device. Network egress is expected for search, but it is a real data-transfer behavior users must understand.
Factores de riesgo
🔑 Variables de entorno (32)
main_claw.py:35 main_claw.py:41 main_claw.py:42 manifest.json:99 README.md:21 README.md:24 README.md:27 README.md:30 README.md:37 README.md:40 requirements.txt:2 scripts/adapters/adapter.py:28 scripts/adapters/adapter.py:73 scripts/adapters/adapter.py:77 scripts/search_core.py:19 scripts/search_core.py:24 scripts/search_core.py:50 scripts/search_core.py:19 scripts/search_core.py:36 scripts/search_core.py:77 scripts/search_core.py:78 scripts/search_core.py:91 scripts/search_core.py:92 scripts/search_core.py:119 skill.json:6 skill.json:7 SKILL.md:46 SKILL.md:49 SKILL.md:52 SKILL.md:55 SKILL.md:62 SKILL.md:68
🌐 Acceso a red (17)
manifest.json:7 manifest.json:57 manifest.json:107 package.json:22 package.json:24 README.md:12 README.md:15 scripts/search_core.py:9 scripts/search_core.py:112 scripts/search_core.py:123 scripts/search_core.py:14 scripts/search_core.py:71 scripts/search_core.py:118 SKILL.md:12 SKILL.md:38 SKILL.md:46 SKILL.md:52
📁 Acceso al sistema de archivos (4)
Auditado por: codex
30 jun 2026, 11:29
Static analysis found many env, network, filesystem, and external command indicators. Review confirms a legitimate community web search client, not malicious code, but it sends queries and an API key to an external endpoint and reads local credential files. The critical heuristic is dismissed as a false positive for intent, with publication allowed only with clear credential and network warnings.
12
Archivos escaneados
770
Líneas analizadas
9
Elementos de revisión
2
Falsos positivos ignorados
Preocupaciones de seguridad confirmadas (1)
Medio
Documentation Encourages Sharing API Key With Assistant
The documentation tells users to ask an AI assistant to configure QRYMA_API_KEY and includes example prompts containing the key placeholder. This can train users to paste real credentials into chat.
The wording and examples are explicit. The risk is documentation-driven credential exposure, not hidden code execution.
Elementos de revisión de capacidades (4)
Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.
Medio
Credential-Bearing Network Request
The search client posts user queries to the configured endpoint and includes the Qryma API key in the X-Api-Key header. This is expected for the skill, but users must understand that query text and credentials leave the local environment.
The code directly builds a POST request to self.endpoint and attaches self.api_key as X-Api-Key. The behavior is semantically legitimate for a search API client, so it is medium rather than high risk.
Medio
Configurable Endpoint Can Receive API Key and Queries
The endpoint can be loaded from QRYMA_ENDPOINT or a .env file before network requests are made. If configuration is tampered with, the API key and search queries could be sent to an unintended server.
Endpoint override behavior is explicit in code and is later used as the request target. Legitimate customization is possible, but the trust boundary is real.
Medio
Local Credential File Reads
The skill reads QRYMA_API_KEY and QRYMA_ENDPOINT from ~/.qryma/.env or .env. This is normal configuration handling, but it accesses hidden home-directory files and local environment files.
The file paths and parsing logic are directly visible. The reads are scoped to Qryma variables, which lowers risk but still requires disclosure.
Bajo
Shell Command Detections Are Documentation Examples
Static analysis flagged shell-like command blocks in README.md and SKILL.md. These are installation and usage examples, not commands executed automatically by the skill.
The suspicious text appears inside Markdown fenced examples. No evidence found that the skill executes these command strings automatically.
Falsos positivos estáticos ignorados (2)
Estas coincidencias estáticas fueron descartadas por la revisión semántica o coincidieron con tokens solo de esquema, por lo que se muestran por transparencia, pero no afectan la puntuación de calidad.
Bajo
Weak Cryptography Detections Are False Positives
Static analysis reported weak cryptographic algorithms across metadata and documentation. Review found descriptions, options, and Markdown format references, with no cryptographic implementation in the cited files.
The cited lines contain metadata, argument descriptions, or output format text. No evidence found of MD5, SHA1, or another weak cryptographic primitive being used.
Bajo
Hardcoded URLs Are Expected Service and Documentation Links
Hardcoded URL findings point to the Qryma service endpoint, project homepage, repository, or language documentation. These are expected for a web search integration.
The URLs match the advertised service or supporting documentation. No evidence found of unrelated exfiltration domains.
Factores de riesgo
🔑 Variables de entorno (10)
🌐 Acceso a red (9)
⚙️ Comandos externos (5)
📁 Acceso al sistema de archivos (6)
Patrones detectados
Network Request With Secret HeaderEndpoint Override From Environment or .envHidden Home Directory Configuration Access
Auditado por: codex
2 abr 2026, 02:51
Static analysis detected 136 potential security issues across 12 files. After evaluation, all findings are false positives related to legitimate skill functionality: environment variable access for API key authentication, network requests to official Qryma API endpoints, shell commands in documentation examples, and configuration file reading. No evidence of malicious intent, credential exfiltration, or data theft. The skill follows standard patterns for API-based search tools with proper error handling.
12
Archivos escaneados
770
Líneas analizadas
3
Elementos de revisión
0
Falsos positivos ignorados
Preocupaciones de seguridad confirmadas (1)
Bajo
Shell Commands in Documentation
Static scanner detected shell backtick patterns in README.md and SKILL.md files. These are documentation code examples within markdown code blocks, not executable code. Lines 10-72 in README.md and lines 17-67 in SKILL.md contain example commands for users to run manually.
Direct evidence these are markdown code blocks, not executable code. The backticks are part of markdown syntax for code examples.
Factores de riesgo
🔑 Variables de entorno (5)
🌐 Acceso a red (2)
Auditado por: claude
26 mar 2026, 02:31
All 112 static analyzer findings were evaluated and determined to be FALSE POSITIVES. The env_access patterns are legitimate API key configuration. Network calls go to the documented Qryma API endpoint. External command findings are markdown documentation examples, not actual shell execution. Filesystem access targets standard configuration directories. No malicious patterns, code execution, data exfiltration, or obfuscation detected. The skill is safe for publication.
12
Archivos escaneados
727
Líneas analizadas
4
Elementos de revisión
0
Falsos positivos ignorados
Elementos de revisión de capacidades (1)
Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.
Bajo
Configuration File Access
The skill reads API configuration from ~/.qryma/.env and .env files. This is standard practice for API key management but involves filesystem access.
Factores de riesgo
🔑 Variables de entorno (3)
🌐 Acceso a red (2)
📁 Acceso al sistema de archivos (2)
Auditado por: claude
25 mar 2026, 09:16
Static analysis flagged 112 potential issues, but most are false positives. Environment variable access (QRYMA_API_KEY, QRYMA_ENDPOINT) is legitimate configuration for an API-based skill. Network calls target the documented Qryma API endpoint. External command findings are documentation examples in markdown, not executable code. Filesystem access is limited to standard config file locations (~/.qryma/.env). No malicious patterns detected.
12
Archivos escaneados
723
Líneas analizadas
5
Elementos de revisión
0
Falsos positivos ignorados
Elementos de revisión de capacidades (2)
Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.
Bajo
Environment Variable Access for API Configuration
Skill reads QRYMA_API_KEY and QRYMA_ENDPOINT from environment variables and ~/.qryma/.env config file. This is expected behavior for an API-based skill but requires user to trust the Qryma service.
Direct evidence of standard API key loading pattern from environment variables and config file - legitimate for API client but requires trust in external service.
Bajo
Network Requests to External API
Skill makes HTTPS POST requests to https://search.qryma.com/api/web endpoint, sending user search queries. This is core functionality but requires trust in the Qryma service with user data.
Clear evidence of network requests to documented Qryma API endpoint - expected behavior for search skill but data leaves user's environment.
Factores de riesgo
🔑 Variables de entorno (6)
🌐 Acceso a red (2)
📁 Acceso al sistema de archivos (2)
Auditado por: claude
25 mar 2026, 08:14
Static analyzer flagged 112 patterns but evaluation shows these are false positives. Environment variable access (QRYMA_API_KEY, QRYMA_ENDPOINT) is standard configuration for API-based tools. Network requests to search.qryma.com represent core functionality. Hidden file access (~/.qryma/.env) is legitimate config storage. No malicious patterns, credential exfiltration, or obfuscation detected. The skill is a transparent web search tool requiring user-provided API credentials.
12
Archivos escaneados
723
Líneas analizadas
7
Elementos de revisión
0
Falsos positivos ignorados
Elementos de revisión de capacidades (4)
Estas son capacidades locales reales que pueden esperarse para esta habilidad, por lo que requieren revisión, pero no se cuentan como comportamiento malicioso confirmado.
Medio
Configuration File in Home Directory
The skill reads configuration from ~/.qryma/.env which is a hidden directory in the user's home folder. While this is standard practice for CLI tools, it represents elevated filesystem access.
Pattern confirms hidden file access but purpose is legitimate configuration storage, not data exfiltration.
Bajo
Environment Variable Access for API Configuration
The skill reads QRYMA_API_KEY and QRYMA_ENDPOINT from environment variables. This is standard practice but represents access to sensitive configuration.
Confirmed environment variable access for API credentials - expected behavior for API-based tools.
Bajo
HTTPS Network Requests to External API
The skill makes POST requests to https://search.qryma.com/api/web to perform searches. This is core functionality but represents network access.
Confirmed network requests to documented API endpoint - legitimate search functionality.
Bajo
Hardcoded API Endpoint URL
Default endpoint URL is hardcoded in the source code. This is acceptable for a skill designed to work with a specific service.
Hardcoded URL to official Qryma service - expected for vendor-specific integration.
Factores de riesgo
🔑 Variables de entorno (4)
🌐 Acceso a red (2)
📁 Acceso al sistema de archivos (2)
Auditado por: claude