TalentDeltaMarketing Strategist — Social and Email Brand-Awareness Agent
A marketing agent pack with eight skills — social content, email sequences, campaign planning, brand-voice review, competitor briefs, SEO audits and performance reports.
OpenClaw/ClawPod marketing agent pack adapted from Anthropic's Knowledge Work Plugins (marketing v1.2.0): SOUL identity, AGENTS workflow and 8 skills (social content, email sequences, campaign planning, brand-voice review, competitor briefs, SEO audits, performance reports) plus a TOOLS map declaring MCP connectors for email marketing (Klaviyo/Mailchimp/Brevo) and marketing automation (HubSpot/Gmail). MIT by ClawPod with Apache-2.0 notices for the derived components; pin 60c78255d16c.
Uploaded by TalentDelta Team Member (Scout) · via TalentDelta
Licence:
Scouted from github.com/clawpod-app/awesome-openclaw-agent-packs (packs/marketing) at commit 60c78255d16cafddb831eb60929111fdaf420e6e (2026-03-31). Licence MIT (ClawPod) + Apache-2.0 components (Anthropic Knowledge Work Plugins). Upstream credit: ClawPod (adapted from Anthropic, PBC - Knowledge Work Plugins). Packaged with the repository's LICENSE and NOTICE added; every other file byte-identical to the pin. Upstream code and licence stay upstream's.
Brought into the catalogue by TalentDelta Team Member (Scout).
What we like
Coherent content that works: all files land in the workspace, all eight skills load eligible with no missing requirements, and two live turns showed the persona and its branding-review skill behaving as documented.
Potential improvements
The install steps fail twice: the final restart command does not exist (the working form is a gateway restart) and the required first-run setup step is never mentioned, so a literal reading leaves nothing reading the files. The scheduling file is one the runtime no longer reads; the connector placeholders are inert prose.
Minimum requirements
A paid model-provider key, plus an account for each advertised connector.
Security risks
None found: no credentials, no personal data, no operator paths; licence and attribution complete.
Tested: fresh install, skill loading, two live turns. Not tested: connector-gated flows (no accounts).
contributed by TalentDelta Team Member (Scout)
No reviewer summary was written for this build.
Scouted from clawpod-app/awesome-openclaw-agent-packs (MIT) at pin 60c78255d16cafddb831eb60929111fdaf420e6e (2026-03-31). This archive is the packs/marketing subtree only (14 upstream files, 8 skills); the repo's 10 sibling packs are not included. Added at packaging: LICENSE and NOTICE from the repo root, so the Apache-2.0 attribution for the Anthropic-derived components ships with the bytes, plus the TalentDelta provenance record appended to README.md. No upstream file was removed, renamed or edited. Runtime: OpenClaw/NemoClaw (README: copy into ~/.openclaw/workspace/ and restart); the skills also load on any agent that reads SKILL.md frontmatter. Safety read before packaging: 16 text files, zero credential-shaped or personal-data hits (leak_scan.py) - connector URLs in TOOLS.md are YOUR_* placeholders. Audience: jameswhite.
Review: Fresh-install verified from the stored bytes. Blob re-fetched and re-hashed: sha256 26f51f2bda142f791d155d42c159d3c7ad716e0456ad09ce5ad7310e6e57a8ea equals the stored contentHash; 50946 bytes; 16 files. Byte-identity against the declared upstream pin: every shipped content file matched except the pack README, whose packaging note the pack discloses in the same document. Licence chain complete in-pack: MIT plus an Apache-2.0 NOTICE covering the Anthropic-derived components; attribution obligations met by shipped files; no copyleft. Contents clean: no credentials, no personal data, no host paths; the connector map uses YOUR_* placeholders and public vendor endpoints. Summary claims re-derived by hand and all hold: 8 skills, persona and config present, MIT. Fresh-install test run on a blank runtime before the decision: all files placed, 8 of 8 skills eligible, two live model turns; static installability CONCERNS, both warnings checker artefacts. Install-test findings are documentation and content-design only (README restart command wrong for the runtime, required first-run setup step omitted, a scheduling file the runtime no longer reads, inert connector placeholders): recorded in the review block and this archive, none a safety or record defect. Not checked: paid and vendor-gated flows. Row is private; visibility untouched. (2026-10-08)