Cursor skills

Personal Agent Skills for this machine. Each skill lives in its own directory with a SKILL.md (and optional scripts/). Credentials stay outside this repo (e.g. ~/.atlassian_config, ~/.gitlab_readonly_config).

Skills

Skill	Description
confluence-reader	Read Confluence pages and spaces (REST v1/v2), CQL search. Uses ~/.atlassian_config. Read-only.
confluence-writer	Publish Markdown to Confluence via md2conf (preferred), optional md_to_confluence_storage.py for a lightweight Markdown subset, or raw storage HTML via curl. Same credentials as confluence-reader; writes only with explicit user authorization.
gitlab-reader	Read-only GitLab API v4: projects, repo files, code search, MRs, pipelines/job logs. Uses ~/.gitlab_readonly_config.
grafana-log-reader	Query Grafana Loki for logs via LogQL through the datasource proxy. Accepts datasource by name or ID. Saves output to temp files for agent analysis. Uses ~/.grafana_config.
grill-me	Interview the user relentlessly about a plan or design until reaching shared understanding. Credit: mattpocock/skills (MIT)
jira-reader	Read Jira issues, search with JQL (REST v3), fetch epic snapshots. Uses ~/.atlassian_config. Read-only.
skill-maintainer	Workflow to stage, commit, and push changes in this repo, and to keep this README aligned with the skills list.
write-a-skill	Create new agent skills with proper structure, progressive disclosure, and bundled resources. Credit: mattpocock/skills (MIT)

Project-local skills and agents (versioned with their respective projects, not this repo):

Skill	Location	Description
jira-writer	~/workspace/NDI/.cursor/skills/	Create and update user stories in Jira under a parent epic. Requires explicit user approval. Uses ~/.atlassian_config.
planning	.cursor/agents/planning.md (epic-planning repo)	Cursor subagent for epic planning workflows.

Credential files

Credentials stay outside this repo. Create the files below and never commit them.

~/.atlassian_config

Used by: jira-reader (including fetch-jira-epic.sh), confluence-reader, confluence-writer.

# Jira + Confluence (core — used by all Atlassian skills)
export JIRA_BASE_URL="https://<org>.atlassian.net"
export CONFLUENCE_URL="https://<org>.atlassian.net/wiki"
export JIRA_EMAIL="<your-email>"
export JIRA_API_TOKEN="<your-atlassian-api-token>"

# md2conf aliases (used by confluence-writer / markdown-to-confluence CLI)
export CONFLUENCE_DOMAIN="<org>.atlassian.net"
export CONFLUENCE_PATH="/wiki/"
export CONFLUENCE_USER_NAME="${JIRA_EMAIL}"
export CONFLUENCE_API_KEY="${JIRA_API_TOKEN}"

# Personal Confluence space (used by confluence-writer scripts)
export CONFLUENCE_SPACE_KEY="~<your-account-id>"
export CONFLUENCE_HOMEPAGE_ID="<numeric-page-id>"

Generate an API token at https://id.atlassian.com/manage-profile/security/api-tokens.

~/.gitlab_readonly_config

Used by: gitlab-reader.

export GITLAB_BASE_PROJECT_URL="https://<gitlab-host>/api/v4/projects/<project-id>"
export GITLAB_READONLY_TOKEN="<personal-access-token>"

Create a read_api scoped token in GitLab → User Settings → Access Tokens.

~/.grafana_config

Used by: grafana-log-reader.

export GRAFANA_URL="https://<your-grafana-instance>"
export GRAFANA_TOKEN="<grafana-service-account-token>"

Create a Viewer service account at Grafana → Administration → Service Accounts, then generate a token.

Layout

~/.cursor/skills/
├── README.md
├── confluence-reader/
├── confluence-writer/
├── gitlab-reader/
├── grafana-log-reader/
├── grill-me/
├── jira-reader/
├── skill-maintainer/
└── write-a-skill/