Background
FM-Agent currently produces individual analysis and bug reports, but users do not have a single result page for discovering, reviewing, and navigating those reports. As the number of findings grows, locating a specific result by status, source file, function, or code location becomes difficult.
User story
As an FM-Agent user, I want a user-facing report index so that I can quickly understand the run results, find relevant findings, and open the underlying report without browsing the output directory manually.
Proposed result
Generate one of the following after a run:
- a self-contained
report.html result page, or
- an interactive TUI result page with equivalent functionality.
The generated result should work from the run artifacts and should not require additional LLM calls.
Required report fields
Each item in the report list must show:
- report title
- status
- source file
- function name
- code location (line number or range when available)
- link or navigation action to the detailed report
Required interactions
- Search across report title, file path, function name, and code location.
- Filter by at least status and source file.
- Sort by useful fields such as status, file, function, and code location.
- Allow users to expand and collapse technical details so the default list remains easy to scan.
- Preserve a clear way to reset active search, filters, and sorting.
Acceptance criteria
Notes
For an HTML implementation, links may point to the existing Markdown/HTML report artifacts, and client-side search/filter/sort is sufficient. For a TUI implementation, selecting an item should open or preview the same detailed report content.
Background
FM-Agent currently produces individual analysis and bug reports, but users do not have a single result page for discovering, reviewing, and navigating those reports. As the number of findings grows, locating a specific result by status, source file, function, or code location becomes difficult.
User story
As an FM-Agent user, I want a user-facing report index so that I can quickly understand the run results, find relevant findings, and open the underlying report without browsing the output directory manually.
Proposed result
Generate one of the following after a run:
report.htmlresult page, orThe generated result should work from the run artifacts and should not require additional LLM calls.
Required report fields
Each item in the report list must show:
Required interactions
Acceptance criteria
report.htmlor exposes an interactive TUI results view.Notes
For an HTML implementation, links may point to the existing Markdown/HTML report artifacts, and client-side search/filter/sort is sufficient. For a TUI implementation, selecting an item should open or preview the same detailed report content.