srekit task¶

Generate an investigation log — a structured artifact for tracking the hypothesis-and-evidence trail when you're hunting a tail-latency spike, a flaky test, or any open-ended SRE puzzle. Hidden alias: srekit sretask (kept for migration from gch sretask).

Synopsis¶

srekit task --title TITLE [flags]

Flags¶

Flag	Required	Description
`--title`	yes	Subject of the investigation; used in the H1 and the default filename
`--path DIR`	no	Directory to write into (default: current dir)

Plus the shared output flags: --out, --stdout, --force, --dry-run, --json.

Default filename¶

If you pass neither --out nor --stdout, srekit writes to <path>/investigation-<slug>.md (lowercased, slug-cleaned).

Examples¶

Quick scratch into stdout:

srekit task --title "Tail latency on api-gw" --stdout

Write into a specific directory:

srekit task --title "Tail latency on api-gw" --path ./tasks
# → ./tasks/investigation-tail-latency-on-api-gw.md

Pipe into jq to grab the generated UUID:

srekit task --title "Tail latency on api-gw" --json | jq -r '.meta.id'

Template shape¶

task ships as a v1 YAML artifact (internal/tmpl/templates/task.yaml) — frontmatter (id, creation_date, modification_date, type: investigation, title, tags), H1, meta_bullets, and six sections: context, hypothesis, evidence, findings, action_items, references. Their headings render bilingually — Russian, with the English term in parentheses; srekit task -T X --json | jq -r '.sections[].title' prints them as they appear in the document. Template expressions inside the YAML reference .Meta.<Field>; the available fields are ID, Title, CreationDate, ModificationDate.