DF star-search
Use when asked to search the web, find online information, research topics, get news, look up Chinese content, or check finance / tech news. **v20.42 - LangChain + Dify + 5 分钟安装体验**: 新增 LangChain Tool 适配器 (`integrations/langchain/star_search_tool.py`) + Dify plugin (`integrations/dify/manifest.yaml`), 一键安装脚本 `install.sh`, `.env.example` 模板, 5 分钟快速开始. **v20.41 基础**: English coverage + open scholarly sources (pure-English queries auto-route to Bing HTTP backend; OpenAlex + CrossRef merge). 16 plus engines, intent understanding, observable per-call metrics. The public service exposes standard MCP (4 tools) plus JSON-RPC and SSE. v20 series highlights: sub-second SSE streaming, multi-turn dialog, 4 output formats, Prometheus monitoring, semantic search, AI orchestration layer (intent classification, entity card, cross-source verification), bot-protection workarounds, and a 4-stage end-to-end pipeline that defaults to LLM answer + auto-fetched snippets. 16 plus engines, intent understanding, and observable per-call metrics.
As a process F 35/100 · Will not run — References files that are not bundled: references/cloudflare-bot-protection.md, scripts/test_intent_108.py, references/v20-40-intent-strat-rule-pitfalls.md
What is at stake
Medium-severity findings: the skill is probably honest, but read what alarmed the scanner.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The skill contains commands that delete files, rewrite disks or execute code fetched from the network. The agent may run them without asking if it believes the instructions require it.
Replace destructive commands with safe equivalents that ask for confirmation, scope them to one folder, and stop piping curl into a shell: pin a version and a checksum.
Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.
The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".
If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.
How to improve
- Shorten the description to 1024 characters.
- For Hermes the description must be one sentence under 60 characters; move the conditions to a "When to Use" section.
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- The text references files that are not there: add them or drop the references.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 16
✓ No critical or high findings
Medium and low: 16
-
medium Dangerous commands
cmd-privilegeinstall.sh:70Privilege escalation / world-writable permissionssudo cp deploy/star-search.service /etc/systemd/system/
-
medium Exfiltration
net-redirectable-api-keyscripts/answer.py:60Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/deep_research.py:16Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/discover.py:143Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/entity_card.py:23Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/query_rewrite.py:17Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Exfiltration
net-redirectable-api-keyscripts/super_brain.py:21Helper sends the API key to a host configured by an environment variable — the key can be redirected to another serverAPI key + configurable base URL from environment
-
medium Dangerous commands
cmd-privilegeSKILL.md:347Privilege escalation / world-writable permissions# 推荐: 把 server 文件属主改成 ubuntu 用户 (sudo chown -R ubuntu:ubuntu <install-dir>),
-
low Exfiltration
read-dotenvinstall.sh:35Reads a .env filecp .env.example .env
-
low Dangerous commands
cmd-background-processinstall.sh:72Starts a background / autostarted processsudo systemctl enable star-search
-
low Dangerous commands
cmd-privilegescripts/search.py:103Privilege escalation / world-writable permissions (string literal in code, not executed)'--no-sandbox', '--disable-setuid-sandbox',
code literal -
low Secrets in code
secret-labelled-tokenscripts/verify.py:6Labelled token / key literal (vendor format unknown — verify it is not a live credential) (quoted — discussed, not commanded)- 测试 secret_key: 1x00…0AA (始终通过)
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/verify.py:6High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)- 测试 secret_key: 1x00…0AA (始终通过)
quoted -
low Secrets in code
secret-password-literalscripts/verify.py:6Hard-coded password / key literal (may be an example) (quoted — discussed, not commanded)- 测试 secret_key: 1x00…0AA (始终通过)
quoted -
low Secrets in code
secret-high-entropy-tokenscripts/verify.py:19High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)TURNSTILE_SECRET_KEY = os.environ.get('TURNSTILE_SECRET_KEY', '1x00…0AA')quoted -
low Dangerous commands
cmd-privilegeSKILL_EN.md:413Privilege escalation / world-writable permissions (quoted — discussed, not commanded)- **docker container mount data dir**: `user:"0"` + `chmod 777` solve mmap permissions
quoted
Files scanned: 50. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1034 chars, limit 1024 - warning
description-long-hermesdescription is 1034 chars; the Hermes authoring standard requires ≤ 60 (one sentence, ending with a period) - warning
body-longSKILL.md body ≈ 5975 tokens (recommended < 5000); move details to references/ - warning
missing-refreference to a missing file: references/cloudflare-bot-protection.md - warning
missing-refreference to a missing file: scripts/test_intent_108.py - warning
missing-refreference to a missing file: references/v20-40-intent-strat-rule-pitfalls.md
Process rating: all ten parameters 35/100
- 0Tools and files. 3 referenced file(s) missing: references/cloudflare-bot-protection.md, scripts/test_intent_108.py, references/v20-40-intent-strat-rule-pitfalls.md
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 1 mutating operations with no state check
- 50Failures and branches. 0 branches, has a failure section
- 60Consistency. The Hermes dialect needs category and tags
- 70Execution cost. Instruction body is 5975 tokens
- 100Steps. 117 steps
- 100Progress reporting. Reports progress
- low 13 top-level sections: this looks like several domains in one skill
- low The response is described with custom markup (13 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 1034: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -47 reference files, but SKILL.md never points to them: the model will not open them
- -317 of 27 scripts are never mentioned in SKILL.md
- +2Single-language instructions
- +4Structure: 52 headings
- +3Step-by-step instructions: 117 items
- +4Has examples (8 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 22.