The open-source skills that turn your agent into a full PR team.
See the codeThe open-source skills that turn your agent into a full PR team.
Install once. Your agent — Claude Code, Codex, Hermes, OpenClaw — becomes a PR team.
curl -fsSL newsjack.sh | bash
Are you an agent? Check out Getting started
Are you a human? 📋 Copy this prompt to any AI:
help me install https://newsjack.sh from the github repo and setup a daily newsjack monitoring for my company.
Jump to platform-specific setup below for per platform breakdown.
Four problems, separate lanes.
Each capability is a skill. Invoke it by name — /angle-generator, /fact-check —
or just describe what you want and your agent picks the right one.
/newsjack-monitor-setup — build the monitoring profile once: your standing, beat topics, competitors, proof assets, spokespeople, feeds, and optional Slack delivery (local agent only)/newsjack-detector — find newsjacking opportunities: fresh stories you have the standing to jump on before the wave breaks, with finished reports optionally delivered to Slack (see a sample run)/news-search — dated, attributed articles on a topic, company, or competitor: when they launch, raise, or stumble, you know/story-origin-check — verify the story is still fresh: who broke it, who owns it, what oxygen's left/relevance-coarse-filter — cheap high-recall first pass that throws out obvious junk before anything expensive runs/newsjack-triage — route surviving stories by whether you actually have standing: pitch now, watch, or surface as big news/coverage-tracker-setup — declare the keywords you want watched, and what each one actually means (local agent only)/coverage-tracker — Google Alerts-style keyword tracking with LLM filtering for real features, alerting only on genuinely new coverage/angle-generator — turn one update into hooks framed for different beats/headline-generator — headlines and pitch subject lines built from the story's raw facts/meanest-editor — roast your pitch against the rubric editors actually use/crisis-holding — holding statements, journalist Q&A posture, and what not to say, with a hard legal-counsel gate/reactive-comment — triage inbound HARO-style source requests, draft only the real fits/fact-check — extract claims, verify each, flag the shaky ones before you send/journalist-fit-check — will this reporter actually care, or are you spamming?/same-outlet-ranker — rank colleagues at the same publication and send to one, instead of spraying the masthead/voice-extractor — fingerprint your real writing, kill the AI tells/find-journalists — build a fit-checked media list, optionally through Medialyst's approval-gated research workflow/press-clip — turn a live article URL into a branded press-clip PDF: the outlet's own logo and layout kept, ads and clutter stripped, the client's mention highlighted (local agent only)/pr-strategist — opinionated walkthrough if you're not PR-fluent yet: audience first, positioning second, news pegs third, drumbeat over big-bang/pr-calendar — plan six months of source-backed upcoming hooks, with prior-year coverage patterns for angle inspiration/newsworthiness-check — cold read on whether it clears the bar before you act/ai-visibility-writing — audit or fact-preservingly revise supplied copy so AI answers can reuse it, without inventing evidence or promising rankings or citations/build-ai-visibility-panel — to measure it, run this one first: it drives the six below it in order, turning any public URL and description into an evidence-bound prompt list across intent, journey, B0–B5 proximity, aided status, roles, locales, surfaces, and tracking partitions/icp-evidence-analysis — turn a company and market dossier into testable ideal-customer hypotheses, buying roles, triggers, disqualifiers, and named research gaps/buyer-job-intent-analysis — recover the jobs, struggling moments, workarounds, and authentic buyer language from real customer, review, forum, and search evidence/prompt-proximity-architecture — blueprint which prompt cells are required, optional, and prohibited before anyone writes prompt wording/realistic-prompt-generation — write natural prompt variants from a target-blind brief, so the panel isn't quietly written to flatter you/prompt-set-qa — gate the prompt set on provenance, contamination, answer leakage, naturalness, and semantic duplicates, quarantining anything the evidence doesn't support/ai-visibility-panel-design — select QA-approved prompts into a versioned tracking panel with partitions, surfaces, locales, repetitions, weights, and refresh rules, keeping aided and unaided denominators separate and gating on a human instead of inventing an “AI visibility score”Newsjack is a set of open skills — plain-Markdown instructions your agent reads — plus a small open-source CLI. Most skills run anywhere your agent runs. A few reach for a live news index, journalist enrichment, media list research, or locally-saved monitoring state — those work best in a local agent.
| Skill | Claude.ai | ChatGPT | ChatGPT Work | Cowork | Claude Code | Codex | Hermes | OpenClaw | Medialyst |
|---|---|---|---|---|---|---|---|---|---|
| Strategize | |||||||||
| pr-strategist | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| pr-calendar | 🔧 | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| newsworthiness-check | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| AI visibility | |||||||||
| ai-visibility-writing | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| build-ai-visibility-panel | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| icp-evidence-analysis | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| buyer-job-intent-analysis | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| prompt-proximity-architecture | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| realistic-prompt-generation | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| prompt-set-qa | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| ai-visibility-panel-design | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| Act | |||||||||
| angle-generator | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| headline-generator | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| meanest-editor | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| crisis-holding | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| reactive-comment | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| fact-check | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| journalist-fit-check | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| same-outlet-ranker | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| voice-extractor | ⚠️ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| find-journalists | 🔧 | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| press-clip | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | 🔜 |
| Detect | |||||||||
| news-search | ✅ | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| story-origin-check | 🔧 | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| relevance-coarse-filter | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| newsjack-triage | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| newsjack-detector | ⚠️ | ⚠️ | ⚠️ | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔜 |
| newsjack-monitor-setup | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | 🔜 |
| coverage-tracker | ⚠️ | ⚠️ | ⚠️ | ⚠️ | ✅ | ✅ | ✅ | ✅ | 🔜 |
| coverage-tracker-setup | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | 🔜 |
Legend:
newsjack-monitor-setup, coverage-tracker-setup) only save a profile or config and schedule it; press-clip drives a real Chrome/Edge browser via Playwright to render and clip the live page, which chat apps can't do.pr-calendar is setup-required but portable: it works anywhere the agent can use the newsjack CLI or a connected Medialyst MCP/connector. The calendar endpoint is free; login is required only to prevent abuse. Consumer ChatGPT is the exception for now because of the same skill-loading limitation called out above.
Set up your agent:
Technical? One line on macOS / Linux (review install.sh first if you like):
curl -fsSL newsjack.sh | bash
Not technical? 📋 Copy this prompt to any AI:
help me setup https://newsjack.sh
It reads the guide and handles the rest — npm fallback, Windows, credentials — on any platform.
Prefer a video? Watch the Newsjack installation walkthrough.
Install the Newsjack plugin:
elvisun/newsjack and confirm.elvisun/newsjack marketplace, find the newsjack@newsjack
plugin, and click Install.newsjack login and approve the printed Medialyst link. If you don't have an account you can create a free one on the Medialyst signup page.For project setup in a local Claude Code plugin install, start with newsjack auth status, then run newsjack login if it says project tools are not enabled before asking the agent to create or open a Medialyst project, add project memory, or upload project files.
Note: A community marketplace plugin, newsjack@claude-community, is pending review
and will be added soon.
The two always-on monitors run here in ⚠️ Limited Mode — a best-effort, one-shot scan with nothing saved. For saved, scheduled monitoring, use a local agent.
ChatGPT runs Newsjack only in a degraded ⚠️ form, because Skills aren't available on consumer accounts. For the full toolkit, use ChatGPT Work or OpenAI's Codex.
ChatGPT Work has the Skills beta, so Newsjack runs there at the same level as
Cowork: everything except the two setup skills and press-clip, which need a
local agent.
~/.newsjack/
├── bin/newsjack # CLI
└── newsjack/ # managed bundle
skills installed to your detected runtime(s):
newsworthiness-check score whether your news clears the bar
headline-generator headlines and subject lines from raw story facts
find-journalists build small, fit-checked journalist lists
...and more
The npm package bundles the same CLI and skills, with the newsjack command installed on PATH.
Curl-installed Newsjack auto-updates from the latest GitHub Release before each run. Set NEWSJACK_AUTO_UPDATE=0 to disable.
You can also install the CLI by itself using npm:
npm i -g newsjack@latest
Go
43.6%
Python
31.1%
HTML
10.7%
Shell
6.5%
JavaScript
4.9%
CSS
1.2%
TypeScript
1.2%
The open-source skills that turn your agent into a full PR team.
See the codeThe open-source skills that turn your agent into a full PR team.
Install once. Your agent — Claude Code, Codex, Hermes, OpenClaw — becomes a PR team.
curl -fsSL newsjack.sh | bash
Are you an agent? Check out Getting started
Are you a human? 📋 Copy this prompt to any AI:
help me install https://newsjack.sh from the github repo and setup a daily newsjack monitoring for my company.
Jump to platform-specific setup below for per platform breakdown.
Four problems, separate lanes.
Each capability is a skill. Invoke it by name — /angle-generator, /fact-check —
or just describe what you want and your agent picks the right one.
/newsjack-monitor-setup — build the monitoring profile once: your standing, beat topics, competitors, proof assets, spokespeople, feeds, and optional Slack delivery (local agent only)/newsjack-detector — find newsjacking opportunities: fresh stories you have the standing to jump on before the wave breaks, with finished reports optionally delivered to Slack (see a sample run)/news-search — dated, attributed articles on a topic, company, or competitor: when they launch, raise, or stumble, you know/story-origin-check — verify the story is still fresh: who broke it, who owns it, what oxygen's left/relevance-coarse-filter — cheap high-recall first pass that throws out obvious junk before anything expensive runs/newsjack-triage — route surviving stories by whether you actually have standing: pitch now, watch, or surface as big news/coverage-tracker-setup — declare the keywords you want watched, and what each one actually means (local agent only)/coverage-tracker — Google Alerts-style keyword tracking with LLM filtering for real features, alerting only on genuinely new coverage/angle-generator — turn one update into hooks framed for different beats/headline-generator — headlines and pitch subject lines built from the story's raw facts/meanest-editor — roast your pitch against the rubric editors actually use/crisis-holding — holding statements, journalist Q&A posture, and what not to say, with a hard legal-counsel gate/reactive-comment — triage inbound HARO-style source requests, draft only the real fits/fact-check — extract claims, verify each, flag the shaky ones before you send/journalist-fit-check — will this reporter actually care, or are you spamming?/same-outlet-ranker — rank colleagues at the same publication and send to one, instead of spraying the masthead/voice-extractor — fingerprint your real writing, kill the AI tells/find-journalists — build a fit-checked media list, optionally through Medialyst's approval-gated research workflow/press-clip — turn a live article URL into a branded press-clip PDF: the outlet's own logo and layout kept, ads and clutter stripped, the client's mention highlighted (local agent only)/pr-strategist — opinionated walkthrough if you're not PR-fluent yet: audience first, positioning second, news pegs third, drumbeat over big-bang/pr-calendar — plan six months of source-backed upcoming hooks, with prior-year coverage patterns for angle inspiration/newsworthiness-check — cold read on whether it clears the bar before you act/ai-visibility-writing — audit or fact-preservingly revise supplied copy so AI answers can reuse it, without inventing evidence or promising rankings or citations/build-ai-visibility-panel — to measure it, run this one first: it drives the six below it in order, turning any public URL and description into an evidence-bound prompt list across intent, journey, B0–B5 proximity, aided status, roles, locales, surfaces, and tracking partitions/icp-evidence-analysis — turn a company and market dossier into testable ideal-customer hypotheses, buying roles, triggers, disqualifiers, and named research gaps/buyer-job-intent-analysis — recover the jobs, struggling moments, workarounds, and authentic buyer language from real customer, review, forum, and search evidence/prompt-proximity-architecture — blueprint which prompt cells are required, optional, and prohibited before anyone writes prompt wording/realistic-prompt-generation — write natural prompt variants from a target-blind brief, so the panel isn't quietly written to flatter you/prompt-set-qa — gate the prompt set on provenance, contamination, answer leakage, naturalness, and semantic duplicates, quarantining anything the evidence doesn't support/ai-visibility-panel-design — select QA-approved prompts into a versioned tracking panel with partitions, surfaces, locales, repetitions, weights, and refresh rules, keeping aided and unaided denominators separate and gating on a human instead of inventing an “AI visibility score”Newsjack is a set of open skills — plain-Markdown instructions your agent reads — plus a small open-source CLI. Most skills run anywhere your agent runs. A few reach for a live news index, journalist enrichment, media list research, or locally-saved monitoring state — those work best in a local agent.
| Skill | Claude.ai | ChatGPT | ChatGPT Work | Cowork | Claude Code | Codex | Hermes | OpenClaw | Medialyst |
|---|---|---|---|---|---|---|---|---|---|
| Strategize | |||||||||
| pr-strategist | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| pr-calendar | 🔧 | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| newsworthiness-check | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| AI visibility | |||||||||
| ai-visibility-writing | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| build-ai-visibility-panel | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| icp-evidence-analysis | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| buyer-job-intent-analysis | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| prompt-proximity-architecture | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| realistic-prompt-generation | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| prompt-set-qa | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| ai-visibility-panel-design | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| Act | |||||||||
| angle-generator | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| headline-generator | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| meanest-editor | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| crisis-holding | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| reactive-comment | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| fact-check | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| journalist-fit-check | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| same-outlet-ranker | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| voice-extractor | ⚠️ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| find-journalists | 🔧 | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| press-clip | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | 🔜 |
| Detect | |||||||||
| news-search | ✅ | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| story-origin-check | 🔧 | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | 🔧 | ✅ |
| relevance-coarse-filter | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| newsjack-triage | ✅ | ⚠️ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| newsjack-detector | ⚠️ | ⚠️ | ⚠️ | ⚠️ | 🔧 | 🔧 | 🔧 | 🔧 | 🔜 |
| newsjack-monitor-setup | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | 🔜 |
| coverage-tracker | ⚠️ | ⚠️ | ⚠️ | ⚠️ | ✅ | ✅ | ✅ | ✅ | 🔜 |
| coverage-tracker-setup | ❌ | ❌ | ❌ | ❌ | ✅ | ✅ | ✅ | ✅ | 🔜 |
Legend:
newsjack-monitor-setup, coverage-tracker-setup) only save a profile or config and schedule it; press-clip drives a real Chrome/Edge browser via Playwright to render and clip the live page, which chat apps can't do.pr-calendar is setup-required but portable: it works anywhere the agent can use the newsjack CLI or a connected Medialyst MCP/connector. The calendar endpoint is free; login is required only to prevent abuse. Consumer ChatGPT is the exception for now because of the same skill-loading limitation called out above.
Set up your agent:
Technical? One line on macOS / Linux (review install.sh first if you like):
curl -fsSL newsjack.sh | bash
Not technical? 📋 Copy this prompt to any AI:
help me setup https://newsjack.sh
It reads the guide and handles the rest — npm fallback, Windows, credentials — on any platform.
Prefer a video? Watch the Newsjack installation walkthrough.
Install the Newsjack plugin:
elvisun/newsjack and confirm.elvisun/newsjack marketplace, find the newsjack@newsjack
plugin, and click Install.newsjack login and approve the printed Medialyst link. If you don't have an account you can create a free one on the Medialyst signup page.For project setup in a local Claude Code plugin install, start with newsjack auth status, then run newsjack login if it says project tools are not enabled before asking the agent to create or open a Medialyst project, add project memory, or upload project files.
Note: A community marketplace plugin, newsjack@claude-community, is pending review
and will be added soon.
The two always-on monitors run here in ⚠️ Limited Mode — a best-effort, one-shot scan with nothing saved. For saved, scheduled monitoring, use a local agent.
ChatGPT runs Newsjack only in a degraded ⚠️ form, because Skills aren't available on consumer accounts. For the full toolkit, use ChatGPT Work or OpenAI's Codex.
ChatGPT Work has the Skills beta, so Newsjack runs there at the same level as
Cowork: everything except the two setup skills and press-clip, which need a
local agent.
~/.newsjack/
├── bin/newsjack # CLI
└── newsjack/ # managed bundle
skills installed to your detected runtime(s):
newsworthiness-check score whether your news clears the bar
headline-generator headlines and subject lines from raw story facts
find-journalists build small, fit-checked journalist lists
...and more
The npm package bundles the same CLI and skills, with the newsjack command installed on PATH.
Curl-installed Newsjack auto-updates from the latest GitHub Release before each run. Set NEWSJACK_AUTO_UPDATE=0 to disable.
You can also install the CLI by itself using npm:
npm i -g newsjack@latest
Go
43.6%
Python
31.1%
HTML
10.7%
Shell
6.5%
JavaScript
4.9%
CSS
1.2%
TypeScript
1.2%