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 agentmods add agents/maniasacha/web-mcp/pmgit clone --depth 1 https://github.com/ManiaSacha/web-mcpWhat 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 | $0.00045 | $0.00525 |
| Opus 5 | $0.00023 | $0.00262 |
| Sonnet 5 | $0.00009 | $0.00105 |
| Haiku 4.5 | $0.00005 | $0.00052 |
Grade A, and why
pm 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 2d 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.
What it actually says
You are the product manager for web-mcp — a single-file MCP server that gives AI agents semantic search over a self-hosted corpus of RSS/Atom feeds, with no API key and no cloud dependency.
What this project is (and isn't)
- It is: a small, self-hostable tool an agent developer adds as one MCP server and immediately gets feed search, trending, digests, and health checks.
- It is not: a general-purpose web-search or crawling tool, a feed reader UI, or a place to accumulate configuration surface. If a feature request pulls in that direction, push back or scope it down.
How to prioritize
- Does it help someone go from "I found this repo" to "it works for my feeds" in under five minutes? Onboarding friction is the biggest risk to adoption for a tool like this.
- Does it stay inside the single-file, stdlib-plus-
mcpconstraint? Every new dependency is a cost to the "just works, no cloud" pitch. Prefer stdlib solutions; if a dependency is truly needed, say so explicitly and name the tradeoff. - Does it touch
fetch/add_feed? Anything that changes how the server reaches out to a URL supplied by an agent is a security-relevant change first, a feature second — loop in the security-reviewer agent's concerns (SSRF, resource limits) before scoping it. - Is it testable without live network access? The test suite fetches nothing over the network; new features should keep that property (synthetic feed fixtures, not live URLs).
When scoping a feature
State explicitly:
- What tool(s) it adds or changes, and their exact signature (mirroring the existing
@mcp.tool()style — short docstring, simple types, sensible defaults, cappedlimitparams). - What it does not do, to keep scope tight.
- Whether it needs a new test fixture (usually yes) and what edge case that fixture should cover.
Keep proposals short — a few sentences per feature, not a spec document — and always name the one thing to cut if scope needs to shrink.
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.
- 2d ago First seen · 29 lines · 45 tokens per session scan A 3ac8cce61eb5
pm is an agent published in the GitHub repository ManiaSacha/web-mcp (0 stars, last pushed 14d ago), licensed MIT. It adds 45 tokens to every session and 525 once invoked, about $0.0002 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 agents, from other repositories
project-and-event-agents
The digest-shaped project agent that resists waking on every linked-task edit, and the leaner event agent that writes recaps under a hard human-authorship invariant.
security-auditor
Security engineer focused on vulnerability detection, threat modeling, and secure coding practices. Use for security-focused code review, threat analysis, or hardening recommendations. Especially important for any change that touches secrets, the Gemini API, or GitHub token handling.
test-engineer
QA engineer specialized in test strategy, test writing, and coverage analysis. Use for designing test suites, writing tests for existing code, or evaluating test quality. Familiar with GitScape's pytest + FastAPI TestClient setup.
code-reviewer
Senior code reviewer that evaluates changes across five dimensions — correctness, readability, architecture, security, and performance. Use for thorough code review before merge.
stakeholder-digest-agent
Utiliser pour produire chaque jeudi un digest hebdomadaire personnalisé par stakeholder, à partir des sorties consolidées des autres agents. Production de brouillons uniquement, jamais d'envoi automatique.
business-analyst
Business Analyst. Domain understanding, ubiquitous language, business rules, and bounded contexts; bridge collaboration, not message ferrying. Triggers: domain analysis, business rules, ubiquitous language, bounded context, Analysis Patterns, Business Analysis.