AI skills & plugins
Specialist AI tools for real work.
Purpose-built workflows for Codex and Claude that research carefully, create useful outputs and keep people in control.
African ideas.
Real impact.
The collection
Small tools. Serious leverage.
Install once, then use focused skill commands inside Codex or Claude Code whenever the workflow comes up.
Website Leads Suite
An evidence-backed workflow for finding businesses that may need a better website, preparing review-ready outreach and delivering only explicitly approved campaigns.
- Verified lead research
- Website opportunity audits
- Personalised outreach drafts
- Approval-gated delivery
What it helps with
- Find businesses with no website, broken pages or weak online visibility.
- Create a four-sheet XLSX lead report with evidence, source links and confidence levels.
- Turn approved findings into specific outreach drafts, then send only after explicit approval.
Safety model
Research uses public information only. Outreach is review-first and never sends automatically.
Example prompt
“Find 10 verified website-improvement leads in Harare and export an XLSX report.”
Install in Codex
npx @ciphertechnologies/website-leads-suite codex$website-leads-suite:website-leadsInstall in Claude Code
npx @ciphertechnologies/website-leads-suite claude/website-leads-suite:website-leadsVeriCV
A truthful CV and job-description comparison tool that helps candidates tailor applications without inventing qualifications, experience or skills.
- Job requirement analysis
- Evidence-grounded CV matching
- ATS-friendly suggestions
- Truthful tailoring support
What it helps with
- Compare a CV against a specific job description.
- Identify supported strengths, genuine gaps and role-specific keywords.
- Improve summaries, achievement wording and application focus without exaggeration.
Safety model
It does not fabricate experience, qualifications, job titles, dates, results or skills.
Example prompt
“Compare my CV with this job description. Show only evidence-grounded strengths, gaps and tailoring suggestions.”
Install in Codex
npx @ciphertechnologies/vericv codex$vericv:vericvInstall in Claude Code
npx @ciphertechnologies/vericv claude/vericv:vericvRepo Revive
An evidence-based recovery assistant for broken, unfinished or inherited repositories, helping teams diagnose project health before making careful repairs.
- Repository health assessments
- Recovery planning
- Verified issue separation
- Conservative repair support
What it helps with
- Diagnose broken builds, missing imports, dependency issues and incomplete setup instructions.
- Assess unfinished features, placeholder code, unsafe configuration and outdated structure.
- Create a recovery plan before repairs, then make small authorised fixes based on verified findings.
Safety model
Diagnosis and planning come first. It avoids broad rewrites and separates verified issues from uncertain findings.
Example prompt
“Diagnose this repository and create an evidence-backed rescue report.”
Install in Codex
npx @ciphertechnologies/repo-revive codex$repo-revive:repo-reviveInstall in Claude Code
npx @ciphertechnologies/repo-revive claude/repo-revive:repo-reviveFaultfinder
An evidence-grounded debugging assistant that explains errors, triages logs and separates root causes from cascade failures before suggesting safe fixes.
- Error explanation
- Build-log triage
- Stack trace review
- Evidence-supported fixes
What it helps with
- Explain unfamiliar errors from builds, deployments, APIs, databases or runtime crashes.
- Find the first meaningful failure in long logs and separate it from secondary errors.
- Review stack traces and suggest the smallest justified repair options.
Safety model
Conclusions are labelled by evidence level: verified, likely, possible or information needed.
Example prompt
“Read this build log and find the first meaningful failure.”
Install in Codex
npx @ciphertechnologies/faultfinder codex$faultfinder:faultfinderInstall in Claude Code
npx @ciphertechnologies/faultfinder claude/faultfinder:faultfinderHow we build
Useful by design. Safe by default.
Evidence before output
Research workflows preserve sources and separate verified facts from inference.
Human approval at the edge
External actions such as outreach stay behind explicit review and approval gates.
Built around real work
Each skill turns a repeatable business process into a clear, dependable workflow.
Have a repeatable process your team runs every week? We can turn it into a focused AI workflow.
Let's build
Bring us your ambitious idea.
Tell us about your project. We'll reply within 24 hours with a plan and a clear quote.