personalize-message
Skill298 starsChanged 6 months agoMITView on GitHub
Generate a personalized outbound message (LinkedIn DM or cold email) for a single prospect by combining a template with the lead's profile, company, and recent signals. Use when the user says 'personalize this message', 'tailor this outreach to the prospect', 'rewrite this template for [name]', 'make this message more personal', or 'generate a personal opener'. Side-effecting — calls Anthropic and may pull live profile/company context via Firecrawl/Unipile/Crustdata.
- Reads credentials
In this repository37 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- answer-linkedin-commentsSkill
- build-routineSkill
- campaign-dashboardSkill
- campaign-strategySkill
- claap-weekly-recapSkill
- cross-platform-schedulerSkill
- debuggerSkill
- enrich-with-signalsSkill
- find-lookalikesSkill
- fullenrich-content-engagersSkill
- fullenrich-event-attendeesSkill
- fullenrich-network-activationSkill
- fullenrich-plg-reverse-lookupSkill
- icp-importSkill
- import-leadsSkill
- integrate-with-yalcSkill
- launch-linkedin-campaignSkill
- lemlist-campaign-from-icpSkill
- list-adaptersSkill
- lost-deal-revival-agentSkill
- notion-reviewSkill
- predictleads-dashboardSkill
- predictleads-lookalikesSkill
- predictleads-signalsSkill
- prospect-discovery-pipelineSkill
- provider-builderSkill
- qualify-leadsSkill
- research-prospectSkill
- run-competitive-intelSkill
- run-doctorSkill
- scrape-post-engagersSkill
- send-cold-emailSkill
- setupSkill
- show-routineSkill
- track-campaignsSkill
- yalc-orientationSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 4 are distinctive, and 4 are 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.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Use this skill when the user says:
- Distinctive: line 46
Onboarding interruption guard: Run:
- Very specific: line 50
If BLOCKED, stop and tell the user to finish yalc-gtm start first.
- Distinctive: line 67
Any toggles?: Ask only if volunteered:
- Very specific: line 75
If --channel linkedin, ensure a Unipile account is configured.
- Distinctive: line 104
Don't run anything unless the user says yes.
- Very specific: line 108
When personalize exits non-zero (missing key, Anthropic 429, lead not found), paste the stderr unchanged.
- Very specific: line 112
Default channel is email. Switch to linkedin for DM-style copy.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
Personalize Message
I'll wrap personalize. I'll ask for the template + target lead, validate inputs, run the CLI, and surface the personalized variant with reasoning so you can approve before sending.
When This Skill Applies
Use this skill when the user says:
- "personalize this message"
- "tailor this outreach to the prospect"
- "rewrite this template for [name]"
- "make this message more personal"
- "generate a personal opener"
NOT this skill (use qualify-leads instead):
- "score these leads" / "qualify these leads" — that runs the 7-gate pipeline, not per-lead copy generation.
NOT this skill (use launch-linkedin-campaign instead):
- "send a LinkedIn outreach to this list" / "start the outbound" — campaign launch handles bulk personalization + sending. This skill is one lead at a time.
NOT this skill (use send-cold-email instead):
- "send this email to [name]" — sending is downstream. This skill produces the variant; you choose where it goes.
What This Skill Does
- Asks for the message template (or accepts pasted text inline) and the target lead (email, linkedin URL, or full record).
- Validates inputs — template non-empty, lead resolvable.
- Shells out to
npx tsx src/cli/index.ts personalize --template <text> --email <email>(or with--linkedin-url,--first-name, etc.). - Parses the CLI output — variant text, reasoning, cost.
- Renders the variant cleanly with a "regenerate?" affordance.
What This Skill Does NOT
- Send the message. That's
send-cold-email(email) orlaunch-linkedin-campaign(LinkedIn) or your manual paste-and-send. - Generate sequences (DM1 + DM2 + follow-ups). That's
launch-linkedin-campaign --variants <n>. - Modify the underlying template or prompt — only the per-lead substitution.
Pre-flight
- Onboarding interruption guard. Run:
Iftest -f ~/.gtm-os/.in-flight-setup && echo "BLOCKED" || echo "OK"BLOCKED, stop and tell the user to finishyalc-gtm startfirst. - Confirm cwd is
~/Desktop/gtm-os/.
Workflow
Step 0 — Ask for inputs
One question at a time:
- What's the template? A few sentences with placeholders the prospect's profile can fill (e.g.
{{first_name}},{{company}}). If they paste raw text, that's the--templatevalue. - Which lead? Email is the default identifier (
--email <email>). They can also pass:--linkedin-url <url>(alternative to email)--first-name <name>/--last-name <name>/--company <name>(manual override fields)--linkedin-account <id>(multi-account routing)--channel <email|linkedin|any>(default:email)--enrich(pull live profile/company context before generating)
- Any toggles? Ask only if volunteered:
--dry-run(preview without persisting; default behavior)--segment-id <id>(route through a segment-specific brand voice)
Step 1 — Validate
- Template non-empty, plausible (not just
{{first_name}}). - Lead identifier (email, linkedin URL, or first-name+company combo) present.
- If
--channel linkedin, ensure a Unipile account is configured.
Step 2 — Shell out
cd ~/Desktop/gtm-os && set -a && source .env.local && set +a && \
npx tsx src/cli/index.ts personalize --template "<text>" --email "<email>"Per the 0.13.0 benchmark, single-command side-effecting skills shell out unconditionally. Anthropic API calls + optional Firecrawl/Unipile context pulls = side-effecting.
Step 3 — Parse output
personalize emits the personalized variant + reasoning + cost. On non-zero exit, surface stderr verbatim.
Step 4 — Render
Show:
- The personalized variant (verbatim)
- The reasoning trail (why this opener vs. another)
- Token cost + any context the system pulled (Firecrawl page, Unipile profile, Crustdata signals)
Step 5 — Offer follow-ups
"Want me to also: (a) regenerate with a different angle (re-run with a tweaked template)? (b) batch-personalize this template across a result set via the campaign launcher? (c) pipe this into a LinkedIn campaign via
launch-linkedin-campaign?"
Don't run anything unless the user says yes.
Failure surfacing — verbatim
When personalize exits non-zero (missing key, Anthropic 429, lead not found), paste the stderr unchanged.
Notes
- Default channel is
email. Switch tolinkedinfor DM-style copy. --enrichadds latency (~5s for Firecrawl + ~2s for Unipile profile) but produces dramatically more specific openers. Recommend it whenever the lead is high-value.- The CLI's
--dry-runis true by default — nothing gets sent. You always review before manually pasting or piping into a campaign.
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.

