jira-ticket-ingestion
Skill0 starsChanged 51 days agoMITView on GitHub
Pull a Jira ticket into a local dossier (chronological markdown plus every screenshot) before analysing it. Use whenever the user mentions a Jira ticket key such as PROJ-1234, pastes a Jira URL, asks to fetch, ingest, read, look at, summarise or investigate a ticket, or asks anything that requires knowing what a ticket actually says. Also use before any triage, since triage without the dossier and its images is not allowed.
- Reads credentials
- Installs packages
In this repository12 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- smoke-checkSkill
- fusion-instanceSkill
- fusion-referenceSkill
- fusion-suiteSkill
- fusion-testSkill
- hcm-architectureSkill
- hcm-configurationSkill
- hcm-debuggingSkill
- triage-batchSkill
- triageSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 0 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 23
If dist/jira_ingest.js does not exist, run npm install && npm run build once, then retry.
- Very specific: line 24
If the script exits with Missing or invalid Jira credentials, print its instructions to the user and stop.
- Very specific: line 32
Check tickets/{KEY}/ticket.md before ingesting: if it is already there and the user did not ask for a refresh, use it as is.
- Very specific: line 37
Read ticket.md in full, then open every file in images/.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Jira ticket ingestion
Turns a Jira ticket into tickets/{KEY}/: ticket.md (header, description, comments in strict
chronological order, screenshots inline), raw.json (stripped payload), images/ and files/.
Jira access is read-only. Never call anything other than GET against Jira, whatever the user asks for. If the user wants a comment posted or a status changed, tell them to do it in Jira themselves.
Running the ingestion
The plugin root is ${CLAUDE_PLUGIN_ROOT} when the plugin is installed, otherwise the repository
directory you are working in. From there:
node dist/jira_ingest.js PROJ-1234- If
dist/jira_ingest.jsdoes not exist, runnpm install && npm run buildonce, then retry. - If the script exits with
Missing or invalid Jira credentials, print its instructions to the user and stop. Do not attempt to work around the credential check, and never ask the user to paste a token into the conversation: it belongs in the gitignored.envfile. - A dossier that already exists is refreshed by re-running the command. Attachments already on disk
are reused, so a re-run is cheap.
--forcere-downloads them. - Batch mode fills the corpus without analysing it:
node dist/jira_ingest.js --jql "project = PROJ AND updated >= -180d" --limit 50
Check tickets/{KEY}/ticket.md before ingesting: if it is already there and the user did not ask
for a refresh, use it as is.
Reading a dossier
Read ticket.md in full, then open every file in images/.
Screenshots are primary evidence, not decoration. In this corpus the exact error message, the page state and the configuration values usually live in the screenshots, not in the text. Never produce a triage without having viewed every image.
While reading the images, extract:
- the verbatim error text, including the JBO, ORA or REST codes, exactly as displayed
- the page, the URL pattern and whether the UI is Redwood or classic
- the environment banner or pod name if visible
- the configuration values shown on setup screens, which are the only trace of the actual configuration this project has
files/ holds HAR captures, logs and spreadsheets. Open them only when the analysis needs them,
they are usually large.
raw.json is the fallback when something in ticket.md looks wrong or truncated: it holds the
untouched field values, timestamps and attachment metadata.
Signs the dossier is incomplete
- a
> WARNING: unresolved image referenceline means a comment points at an attachment Jira did not return. Say so in the analysis instead of guessing what the image showed. - a section
## Attachments not referenced inlinewith many images usually means the reporter attached evidence without describing it. Those images still count as evidence, open them. (empty description)on an Incident is itself a fact: the ticket was opened without a written problem statement and everything must come from comments and screenshots.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

