signal-watch
Skill1 starsChanged 4 months agoMITView on GitHub
Daily "who do I contact today" sweep — scan for fresh buying signals (new hires, product launches, incidents, renewal windows) and turn them into a prioritized action list. Use when the rep says "signal watch", "who should I reach out to today", "run my morning sweep", or "any new signals".
In this repository14 other files it gives its agents.
- account-briefSkill
- battlecardSkill
- build-listSkill
- dojoSkill
- draft-sequenceSkill
- handle-replySkill
- meeting-prepSkill
- post-callSkill
- reactivateSkill
- setupSkill
- voice-interviewSkill
- what-worksSkill
- work-inboundSkill
- write-emailSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 12
Read ${CLAUDE_PLUGIN_ROOT}/references/icp.md and references/messaging.md if not already in context.
- Distinctive: line 15
If Slack is connected: which channel the daily list should post to (a public channel is the point — visible momentum.
- Distinctive: line 19
Run these searches, restricted to the last 7 days where the tool allows (include the recency in the query — "this week", "announced this …
- Distinctive: line 27
Keep total searches ≤ 10. Signals without a company you could actually contact are trivia — drop them.
- Distinctive: line 35
Before recommending action on a company, check HubSpot: active pipeline → drop it (note as "signal on active account — FYI [owner]" …
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Signal Watch
The morning routine: fresh signals → matched to plays → ranked actions. Target runtime: one coffee.
Before you sweep
Read ${CLAUDE_PLUGIN_ROOT}/references/icp.md and references/messaging.md if not already in context. Ask the rep (first run only, then remember in a signal-watch-config.md file in the working folder):
- Named accounts they're actively working (their "watch list")
- Any vertical focus this week
- If Slack is connected: which channel the daily list should post to (a public channel is the point — visible momentum; suggest confirming the channel choice with your marketing lead)
The sweep (Exa, fresh content)
Run these searches, restricted to the last 7 days where the tool allows (include the recency in the query — "this week", "announced this week", current month/year):
- New-hire trigger ⭐ — new job postings for the role tied to your primary trigger AND "started new role" announcements at companies matching your ICP firmographics. (A relevant new hire needs a solution fast — often the highest-value signal available; see the ICP reference.)
- Primary-trigger launches ⭐ — companies showing evidence of the primary trigger event this week (feeds Hook #2, the primary play)
- Pain in the wild — practitioners or leaders posting publicly about the problem you solve (relevant subreddits, HN, LinkedIn)
- Displacement signals — posts or job listings referencing frustration with the legacy incumbent or a competing tool, or migration away from it (see messaging.md for approved names)
- Watch-list check — one search per named account for anything new (skip if watch list > 10; do top 5 by ICP score)
Keep total searches ≤ 10. Signals without a company you could actually contact are trivia — drop them.
Scoring what comes back
For each signal: which of the trigger events in the ICP reference is it, how fresh, does the company clear Tier 1/2 firmographics at a glance, and is there a findable Champion/Practitioner? Two correlated signals on one company (e.g., a relevant new-hire post + a primary-trigger launch) beats one signal anywhere — rank those first.
CRM check
Before recommending action on a company, check HubSpot: active pipeline → drop it (note as "signal on active account — FYI [owner]" instead, that's still useful intel to pass along).
Output format
# Signal Watch · [date]
## Act today (max 5)
1. **[Company]** — [signal + link, freshness] → Hook #[N], contact [name/title if found]
First-line starter: "[one sentence the rep can build the trigger line from]"
2. ...
## Worth a look (3-5)
[company — signal — why it's not top-5]
## Watch-list updates
[account → anything new, or "quiet"]
## FYI — signals on active-pipeline accounts
[pass to the owning rep / your marketing lead]End with: "Say brief me on [company] for the full research, or write an email to [name] to go straight to copy."
Slack posting (if the Slack connector is available)
After showing the rep the full output, offer to post a condensed version to the configured channel. Post ONLY after they say yes — they may want to trim a name first. The Slack version is the "Act today" list plus a one-line week-context footer, formatted for Slack (no markdown tables):
:sunrise: Signal watch · [date]
1. [Company] — [signal, one line] → [hook name]
...
[N] worth-a-look accounts in the full run · [meetings booked this week so far]Keep it to the top 5 — the channel post is a momentum artifact, not the working document. Never include anything from the "FYI — active pipeline" section in the public post (deal intel stays out of public channels; DM the owning rep instead). If the post fails because the Slack app isn't approved in the workspace yet, tell the rep to ping your admin and continue without posting.
Weekly discipline
On Fridays, add a one-table recap: signals surfaced this week → actioned → meetings created. If signals aren't converting to actions, say so — the search angles need tuning, and the rep should flag it to your admin or marketing lead.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

