Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Set up and conduct media monitoring to track brand mentions, sentiment, and share of voice across news, social, and online channels. Use this skill when the user needs to track what's being said about their brand, monitor competitors' media presence, detect emerging PR issues early, or measure campaign reach — even if they say 'what are people saying about us', 'monitor our brand mentions', 'track competitor PR', or 'set up media alerts'.
.claude/skills/asgard-ai-platform-pr-media-monitoring/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -7% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -3% | 0% |
| case-09 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 39% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 45% | 0% |
IRON LAW: Monitor for Action, Not Just Awareness
Media monitoring that produces weekly reports nobody reads is waste.
Every monitoring setup must have TRIGGER ACTIONS defined: if sentiment
drops below X, alert the PR team. If a competitor launches a campaign,
notify marketing within 24 hours. No triggers = no value.| Dimension | What to Track | Tools | |-----------|-------------|-------| | Volume | Number of mentions over time | Google Alerts, Mention, Meltwater | | Sentiment | Positive / Neutral / Negative ratio | Brandwatch, Talkwalker, manual coding | | Share of Voice | Your mentions vs competitors' mentions | Industry reports, custom dashboards | | Source | Where mentions appear (news, social, forums, blogs) | Platform analytics, social listening tools | | Influencer | Who is talking (reach, authority) | BuzzSumo, social platform analytics | | Topics | What themes/keywords are associated with your brand | Keyword clustering, topic modeling |
Phase 1: Define Keywords
Phase 2: Select Channels
Phase 3: Set Triggers | Trigger | Threshold | Action | |---------|-----------|--------| | Negative sentiment spike | >20% increase in 24hrs | Alert PR manager immediately | | Competitor product launch | Any mention | Notify marketing team | | Influencer mention (>10K followers) | Any mention | Evaluate for engagement opportunity | | Crisis keyword detected | "recall", "lawsuit", "scandal" + brand | Activate crisis protocol |
Phase 4: Reporting Cadence
markdown# Media Monitoring Report: {Brand} — {Period} ## Summary | Metric | Current | Prior Period | Change | |--------|---------|-------------|--------| | Total Mentions | {N} | {N} | {%} | | Sentiment (Pos/Neu/Neg) | {%}/{%}/{%} | ... | ... | | Share of Voice | {%} | {%} | {±%} | ## Top Mentions | Date | Source | Headline/Summary | Sentiment | Reach | |------|--------|-----------------|-----------|-------| | {date} | {outlet} | {summary} | +/0/- | {est. reach} | ## Alerts Triggered | Date | Trigger | Action Taken | |------|---------|-------------| | {date} | {what happened} | {response} | ## Competitor Activity | Competitor | Mentions | Key Activity | |-----------|----------|-------------| | {name} | {N} | {what they did} | ## Recommendations 1. {action based on findings}
references/taiwan-media.md| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 21,550 | 49,007 | +127% | 1 | 1 | 0% | 3,777 | 3,509 | -7% | 0 | 0 | — |
case-02 | fail→fail | 54,026 | 28,224 | -48% | 1 | 1 | 0% | 7,483 | 5,720 | -24% | 0 | 0 | — |
case-03 | fail→pass | 28,848 | 19,520 | -32% | 1 | 1 | 0% | 4,191 | 4,074 | -3% | 0 | 0 | — |
case-04 | pass→pass | 9,552 | 9,506 | -0% | 1 | 1 | 0% | 1,628 | 2,205 | +35% | 0 | 0 | — |
case-05 | pass→pass | 14,172 | 11,457 | -19% | 1 | 1 | 0% | 1,743 | 2,453 | +41% | 0 | 0 | — |
case-06 | pass→pass | 20,786 | 18,332 | -12% | 1 | 1 | 0% | 2,684 | 3,387 | +26% | 0 | 0 | — |
case-07 | pass→pass | 16,774 | 33,840 | +102% | 1 | 1 | 0% | 2,462 | 3,643 | +48% | 0 | 0 | — |
case-08 | pass→pass | 15,558 | 12,070 | -22% | 1 | 1 | 0% | 2,064 | 2,639 | +28% | 0 | 0 | — |
case-09 | fail→pass | 20,765 | 17,764 | -14% | 1 | 1 | 0% | 2,694 | 3,284 | +22% | 0 | 0 | — |
case-10 | fail→pass | 13,999 | 10,403 | -26% | 1 | 1 | 0% | 1,845 | 2,571 | +39% | 0 | 0 | — |
case-11 | pass→pass | 16,113 | 12,220 | -24% | 1 | 1 | 0% | 2,027 | 2,510 | +24% | 0 | 0 | — |
case-12 | pass→pass | 15,654 | 15,944 | +2% | 1 | 1 | 0% | 2,285 | 3,088 | +35% | 0 | 0 | — |
case-13 | pass→pass | 16,553 | 16,887 | +2% | 1 | 1 | 0% | 2,346 | 3,344 | +43% | 0 | 0 | — |
case-14 | fail→pass | 16,186 | 16,529 | +2% | 1 | 1 | 0% | 2,409 | 3,503 | +45% | 0 | 0 | — |
case-15 | fail→pass | 18,933 | 14,649 | -23% | 1 | 1 | 0% | 2,461 | 3,345 | +36% | 0 | 0 | — |
case-16 | pass→pass | 16,217 | 12,185 | -25% | 1 | 1 | 0% | 2,341 | 2,850 | +22% | 0 | 0 | — |
case-17 | fail→pass | 11,709 | 7,285 | -38% | 1 | 1 | 0% | 2,047 | 2,150 | +5% | 0 | 0 | — |
case-18 | pass→pass | 11,233 | 9,278 | -17% | 1 | 1 | 0% | 1,855 | 2,297 | +24% | 0 | 0 | — |
case-19 | fail→pass | 15,825 | 9,211 | -42% | 1 | 1 | 0% | 2,080 | 2,127 | +2% | 0 | 0 | — |
case-20 | fail→pass | 12,575 | 9,756 | -22% | 1 | 1 | 0% | 1,871 | 2,184 | +17% | 0 | 0 | — |
case-21 | fail→pass | 14,051 | 3,795 | -73% | 1 | 1 | 0% | 2,052 | 1,495 | -27% | 0 | 0 | — |
case-22 | fail→pass | 13,484 | 2,174 | -84% | 1 | 1 | 0% | 2,042 | 1,247 | -39% | 0 | 0 | — |
case-23 | pass→pass | 9,581 | 6,331 | -34% | 1 | 1 | 0% | 1,566 | 1,802 | +15% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +48 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.