个人作者
Daniel Rosehill
Focus: Agentic AI, MCP, and workflow automation (esp. voice assisted). Exploring: multiagent use-cases for AI & geopol sims. GH activity: Claude assisted.
- Skills
- 11
- 仓库
- 1
- 职业覆盖
- 1
- 场景覆盖
- 2
- Stars
- ★ 0
Daniel Rosehill 的 Skills 11
vendor-preferences
Record the user's own preferences for contacting a vendor — which channel they prefer for bugs, feature requests or support, channels to avoid, private addresses, and personal contacts at the company (a developer they know, an account manager, a WhatsApp number, someone met in person). Use when the user says "I prefer to email X", "don't post X stuff on Reddit", "I know someone at X", "for X feature requests go to …", or "remember that for X …".
security-disclosure
Disclose a security vulnerability to a software vendor responsibly, through a private channel only — bug-bounty program, security.txt contact, security@ address, or a trusted contact at the company — never a public tracker, forum or social post. Use when the user has found a vulnerability, data exposure, auth bypass, secret leak or other security flaw in someone else's product and wants to report it.
route
Work out where to send something to a software vendor — a bug report, feature request, security vulnerability, support question or general feedback — taking the vendor's own channels and the user's personal preferences and contacts into account. Use when the user asks "where do I report this to X", "how do I request a feature from X", "who do I tell about this bug", or before any of the report, security-disclosure or vendor-specific filing skills.
report
Send a bug report, feature request, support question or feedback to a software vendor through the right channel — email, GitHub issue, forum or subreddit post, contact form, in-app form, or a message to a personal contact at the company — after checking it hasn't already been reported. Use when the user says "report this to X", "send X a feature request", "email the developer about this bug", "tell my contact at X". For security vulnerabilities use security-disclosure instead.
op-fr-tracker
Log 1Password feature requests and threads the user posted or joined, and check them for staff replies or shipped features.
op-fr-search
Search the 1Password Community and release notes for an existing feature request or answer before posting a new one.
op-feature-request
Draft and submit a feature request to 1Password as a Community topic tagged Feature Request, after a duplicate check and approval.
op-community-engage
Engage in the 1Password Community — +1 or add a use case to a feature request, reply to threads, answer questions, browse new topics.
op-bug-report
Report a 1Password bug on the Community — version check against latest release, duplicate search, environment details, tagged topic.
cc-bug-report
File a bug report against Claude Code (anthropics/claude-code on GitHub), after checking that it hasn't already been reported. Use when the user wants to "report a Claude Code bug", "file an issue with Anthropic", "open a bug on claude-code", or describes Claude Code (the CLI, harness, tools, hooks, plugins, MCP handling) misbehaving. Searches open and closed issues first, offers to comment on a match, and files the issue through Anthropic's live bug_report form only when nothing matches. Not for complaints about what the model said; that is a model-behavior report.
add-vendor
Research how a software vendor takes bug reports, feature requests, security reports and support, and record it as structured channel data (vendor.json) so routing works for that vendor. Use when a vendor isn't known yet, when the user says "add X to vendor liaison", "find out how X takes feedback", or when a vendor's channels have changed or died.