Get bias scores for every news source in the Helium database. Returns a list of all sources (active within the last 36 days, with >100 articles analyzed), sorted by avg_social_shares descending. Use this to compare sources, find the most credible outlets, identify politically extreme...
Part of the Helium MCP server. Enforce policies on this tool with Intercept, the open-source MCP proxy.
AI agents call get_all_source_biases to retrieve information from Helium without modifying any data. This is common in research, monitoring, and reporting workflows where the agent needs context before taking action. Because read operations don't change state, they are generally safe to allow without restrictions -- but you may still want rate limits to control API costs.
Even though get_all_source_biases only reads data, uncontrolled read access can leak sensitive information or rack up API costs. An agent caught in a retry loop could make thousands of calls per minute. A rate limit gives you a safety net without blocking legitimate use.
Read-only tools are safe to allow by default. No rate limit needed unless you want to control costs.
tools:
get_all_source_biases:
rules:
- action: allowSee the full Helium policy for all 9 tools.
Agents calling read-class tools like get_all_source_biases have been implicated in these attack patterns. Read the full case and prevention policy for each:
Other tools in the Read risk category across the catalogue. The same policy patterns (rate-limit, allow) apply to each.
Get bias scores for every news source in the Helium database. Returns a list of all sources (active within the last 36 days, with >100 articles analyzed), sorted by avg_social_shares descending. Use this to compare sources, find the most credible outlets, identify politically extreme sources, or build a ranked overview of the media landscape. Each entry contains: - source_name, slug_name, page_url - articles_analyzed: total articles analyzed for this source - avg_social_shares: average social shares per article (proxy for reach/influence) - emotionality_score (0-10): average emotional intensity of the writing - prescriptiveness_score (0-10): how much the source tells readers what to think/do - bias_values: dict mapping classifier key → integer score (-50 to +50 for bipolar, 0 to +50 for unipolar). These keys are identical to what get_bias_from_url returns, so you can compare article-level and source-level scores directly. Political / ideological (bipolar: neg=left pole, pos=right pole): 'liberal conservative bias' neg=liberal, pos=conservative 'libertarian authoritarian bias' neg=libertarian, pos=authoritarian 'dovish hawkish bias' neg=dovish, pos=hawkish 'establishment bias' neg=anti-establishment, pos=pro-establishment Credibility / quality (bipolar): 'overall credibility' neg=uncredible, pos=credible 'integrity bias' neg=low integrity, pos=high integrity 'article intelligence' neg=low intelligence, pos=high intelligence 'delusion bias' neg=truth-seeking, pos=delusional 'objective subjective bias' neg=objective, pos=subjective 'bearish bullish bias' neg=bearish, pos=bullish 'emotional bias' neg=negative tone, pos=positive tone Unipolar bias dimensions (higher = more of that trait): 'objective sensational bias' sensationalism 'opinion bias' opinion vs informative 'descriptive prescriptive bias' prescriptive vs descriptive 'political bias' political content 'fearful bias' fear-based framing 'overconfidence bias' overconfidence 'gossip bias' gossip 'manipulation bias' manipulative framing 'ideological bias' ideological rigidity 'conspiracy bias' conspiracy content 'double standard bias' double standards 'virtue signal bias' virtue signaling 'oversimplification bias' oversimplification 'appeal to authority bias' appeal to authority 'begging the question bias' question-begging 'victimization bias' victimization framing 'terrorism bias' terrorism content 'scapegoat bias' scapegoating 'suicidal empathy bias' suicidal-empathy framing 'cruelty bias' cruelty 'woke bias' woke framing 'written by AI' AI-written likelihood 'immature bias' immaturity 'circular reasoning bias' circular reasoning 'covering the response bias' covering-the-response tactic 'spam bias' spam-like content Tip: use get_source_bias for full narrative descriptions and recent articles on a specific source. Tip: bias_values keys here are identical to those in get_bias_from_url and search_news — compare them directly. Warning: get_source_bias returns bias_scores with emoji-prefixed display keys (e.g. '🔵 Liberal <—> Conservative 🔴') that are NOT interchangeable with the plain-text keys used here. Do not cross-reference them. . It is categorised as a Read tool in the Helium MCP Server, which means it retrieves data without modifying state.
Add a rule in your Intercept YAML policy under the tools section for get_all_source_biases. You can allow, deny, rate-limit, or validate arguments. Then run Intercept as a proxy in front of the Helium MCP server.
get_all_source_biases is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the get_all_source_biases rule in your Intercept policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the Intercept policy for get_all_source_biases. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
get_all_source_biases is provided by the Helium MCP server (conner-m4el/helium-mcp). Intercept sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Open source. One binary. Zero dependencies.
npx -y @policylayer/intercept