| name | cross-source-gap-analysis |
| description | Compare findings across Slack, GitHub, NVIDIA forums, and Outlook to identify alignment gaps, missing coverage, and follow-ups. |
cross-source-gap-analysis
Use this skill when the task is to compare or synthesize information across
multiple sources rather than merely access one source.
When to use
- Compare Slack discussion against GitHub issues or PRs
- Compare Slack or GitHub findings against NVIDIA forum discussion
- Identify missing coverage, inconsistent narratives, or follow-up areas across sources
- Find external email discussions not reflected in internal meeting updates
Inputs
Load the source skills you need first:
slack-channel-summarizerslack-channel-findergithub-readonly-livesource-etl-queryoutlook-email-search
This skill does not define how to access those systems. It defines how to
combine the findings once you have them.
Procedure
1. Gather the minimum useful evidence from each source
Prefer a small, relevant slice from each source over broad collection. For example:
- a recent Slack window for the relevant channel
- live GitHub issues or PRs from
$GITHUB_READONLY_REPO, when current state
matters
- mirrored GitHub discussions, or historical issue/PR mirror data for the repo
or feature area, when the task is about the ETL mirror or discussions
- mirrored NVIDIA forum topics for the
nemoclaw tag scope
- recent emails filtered to the relevant project and date range
2. Normalize what each source is saying
Reduce each source to short bullets such as:
- active topics
- reported problems
- decisions or planned work
- requests for help
- unresolved questions
3. Compare across sources
4. Present the result
A good default structure is:
- scope and time window
- what all sources agree on
- gaps or mismatches
- concrete follow-ups
Keep the comparison grounded in evidence from the sources you actually checked.
Do not invent gaps just because one source had less data available.
Pitfalls
- Do not force a gap-analysis framing when the user only asked for source access
or a plain summary.
- Do not over-collect. A narrow comparison is usually better than an exhaustive scrape.
- Distinguish between "not discussed" and "not observed in the sampled data".
- Distinguish between “not observed in the mirror” and “not present on the live source”.
- Keep live GitHub REST findings separate from source ETL mirror findings when
the scopes or freshness differ.
