Integration guide
Eleven systems each have a dedicated integration guide, published on the wiki and mirrored locally. This skill's job is to find the right one and follow it — not to carry integration knowledge itself.
1. Match the name to a canonical key
| Canonical key | Aliases | Wiki page | Local fallback file |
|---|---|---|---|
superpowers |
Superpowers | Integration-Superpowers |
docs/integrations/fallback-superpowers.md |
gsd |
GSD, Get Shit Done | Integration-GSD |
docs/integrations/fallback-gsd.md |
speckit |
SpecKit, Spec Kit | Integration-SpecKit |
docs/integrations/fallback-speckit.md |
openspec |
OpenSpec | Integration-OpenSpec |
docs/integrations/fallback-openspec.md |
jira |
Jira, Atlassian Jira | Integration-Jira |
docs/integrations/fallback-jira.md |
confluence |
Confluence, Atlassian Confluence | Integration-Confluence |
docs/integrations/fallback-confluence.md |
github |
GitHub, GitHub Issues | Integration-GitHub |
docs/integrations/fallback-github.md |
gitlab |
GitLab | Integration-GitLab |
docs/integrations/fallback-gitlab.md |
azuredevops |
Azure DevOps, ADO, Azure Boards | Integration-AzureDevOps |
docs/integrations/fallback-azuredevops.md |
awscodecatalyst |
AWS CodeCatalyst, CodeCatalyst | Integration-AWSCodeCatalyst |
docs/integrations/fallback-awscodecatalyst.md |
googleclouddevops |
Google Cloud DevOps, GCP, Google Cloud | Integration-GoogleCloudDevOps |
docs/integrations/fallback-googleclouddevops.md |
If the request names something not in this table (e.g. Linear, Notion),
say so and stop rather than guessing a page that doesn't exist — offer to
add one (edit docs/integrations/README.md and register a new fallback
file the same way as the eleven above).
2. Try the live wiki page first
Build the URL from wiki.root_url in .work/config.yml (currently
https://github.com/SpillwaveSolutions/wiki_ticket_sdd/wiki) plus
/<Wiki page> from the table above. WebFetch it. The wiki copy is the
freshest source — per wiki-publish/SKILL.md, pages can be hand-edited
directly in the web UI, so it can be ahead of what's committed here.
Verify the fetch actually returned that page before trusting it. A
GitHub wiki does not 404 on a missing page slug — it silently redirects to
the wiki's Home page with a normal 200 response. Confirm the fetched
content contains a ## Recommended workflow heading and looks like the
requested system's page (not the Home page's "What is this project?"
sections) before treating it as a real hit. If it doesn't match, treat
this exactly like a fetch failure (§3) — do not follow the Home page as if
it were the integration guide.
On a confirmed hit: say "Referring to the official integration guide for
<System>, published at <url>." Then follow that page's Recommended
workflow section exactly — do not improvise around it.
3. Fall back to the local copy on fetch failure
If the fetch fails outright, or succeeds but doesn't verify per §2 (network
error, 404, wrong-page redirect, wiki not yet created), read the local
fallback file instead and say so explicitly: "Wiki page unavailable — using
the bundled local copy (<path>), which may lag the published page." Same
instruction applies: follow its Recommended workflow section.
4. Soft version-staleness check
If the retrieved page states a minimum tool version, and the system has a
real, probeable CLI (gh --version, glab --version, az --version,
aws --version), run it and compare. Warn if older — never block. The
four SDD tools (Superpowers, GSD, SpecKit, OpenSpec) are Claude Code
skills, not installed binaries: there is nothing to introspect, so say
plainly that no version check applies rather than fabricating one.
5. Compose, don't reinvent
For Jira or Confluence specifically: before doing anything with a raw
REST call or CLI, check whether the global jira or confluence skill
(or an Atlassian MCP server) is available and use it instead — it already
owns auth, pagination, and markup conversion. The full instruction lives
in Integration-Jira / Integration-Confluence themselves (Recommended
workflow section); this is only a pointer so you don't skip straight to
raw API calls when a better tool is sitting right there.
6. Maintaining these pages
After editing a file under docs/integrations/, register it (or refresh
its ledger entry) and let the existing publish pipeline carry it, exactly
like any other opted-in doc:
bin/worklog wiki-add docs/integrations/fallback-<key>.md --key integrations/<key> --title "Integration-<Name>"
Then run the wiki-publish skill as usual — its hash-compare skip logic
picks up the change on the next run. There is no separate publish path for
these files and none should be built; wiki-add plus the ordinary
wiki-publish flow is the whole mechanism (see wiki-publish/SKILL.md
§4). Adding a twelfth system means the same two steps: a new row in the
table above, a new fallback file, one wiki-add call.