Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add ManiaSacha/web-mcp --skill feed-digestgit clone --depth 1 https://github.com/ManiaSacha/web-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/maniasacha/web-mcp/feed-digest)<a href="https://agentmods.dev/skills/maniasacha/web-mcp/feed-digest"><img src="https://agentmods.dev/badge/skills/maniasacha/web-mcp/feed-digest/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/maniasacha/web-mcp/feed-digest"><img src="https://agentmods.dev/badge/skills/maniasacha/web-mcp/feed-digest.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00053 | $0.00614 |
| Opus 5 | $0.00026 | $0.00307 |
| Sonnet 5 | $0.00011 | $0.00123 |
| Haiku 4.5 | $0.00005 | $0.00061 |
Grade A, and why
feed-digest scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 30 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Feed digest
Turns web-mcp's raw tools (search, trending, digest, recent) into a synthesized answer for the user, rather than just relaying tool output verbatim.
Workflow
- Scope the question. If the user gave a topic ("what's new with rocket launches"), use
search(query, max_days=...)with a query built from their words, notdigest()— digest is for "what's happening across everything," search is for "what's happening on this topic." - Scope the time window. Default to 24h for "today"/"this week" style requests map to
hours=168. If the user doesn't say, ask or default to 24h and say so explicitly in the answer. - Check trending first for broad requests. For an open-ended "what's happening" ask, call
trending(hours=...)to see the dominant terms, thendigest(hours=...)for the grouped article list — use the trending terms to decide what to lead with in your summary rather than just listing feeds in whatever orderdigestreturns them. - Synthesize, don't relay. The tools return markdown lists of links; don't just paste that back as the whole answer. Write 3-6 sentences (or a short set of bullets grouped by theme, not by source feed) that actually say what happened, then include the most relevant links as supporting citations.
- Say when there's nothing. If
digest()/search()come back empty or with "No recent articles," say that plainly — don't pad a non-answer into paragraphs.
Handling multiple topics
If the user asks about several distinct topics in one digest request, run separate search() calls per topic rather than one broad query — TF-IDF search degrades when the query mixes unrelated terms, since it's a bag-of-words match, not semantic disambiguation between topics.
Freshness
If the last successful refresh for the relevant feeds looks stale (check source_health() if the user seems to expect very recent coverage and results look thin), mention that the digest reflects the last successful fetch rather than guaranteeing real-time coverage, and offer to call refresh() before re-running the digest.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 30 lines · 53 tokens per session scan A 296b9252c545
feed-digest is a skill published in the GitHub repository ManiaSacha/web-mcp (0 stars, last pushed 20d ago), licensed MIT. It adds 53 tokens to every session and 614 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
rss-digest
Subscribe to RSS/Atom feeds with keyword filters and run digests delivered to Telegram. Use when subscribing to feeds, checking which feeds run, running a digest manually, or adjusting feed keywords.
feishu-memory-recall
Cross-group memory, search, and event sharing for OpenClaw Feishu agents.
building-threat-intelligence-feed-integration
Builds automated threat intelligence feed integration pipelines connecting STIX/TAXII feeds, open-source threat intel, and commercial TI platforms into SIEM and security tools for real-time IOC matching and alerting. Use when SOC teams need to operationalize threat intelligence by automating feed ingestion…
knowledge-compile
A knowledge-organizing tool that turns a folder of scattered Markdown notes, articles, and captured posts into an Obsidian-style wiki. It creates an index, topic pages with links back to the originals, and a short daily digest.
support-digest
A local tool that reads exported customer-support tickets in JSON or CSV and produces a daily summary after hiding personal information. It does not fetch data from support systems or call outside APIs.
stream-flutter
Build and integrate Stream Chat, Video, and Feeds in Flutter apps. Use for Flutter/Dart project work with Stream package setup, auth wiring, and widget blueprints. Supports streamchatflutter (pre-built Chat UI), streamchatfluttercore (custom Chat UI), streamvideoflutter (Video calling and livestreaming), and…