forgeplan discover show

forgeplan discover show <SESSION_ID> prints the current state of a brownfield discovery session — the protocol stage it’s in, how many findings the agent has submitted, and a categorized summary of what’s been captured so far.

Use it during a run to watch progress, and before discover complete to sanity-check coverage.

When to use

Mid-run — the agent has been appending findings and you want to see the picture so far.
Before discover complete — verify the session actually covered the areas you care about.
While debugging — confirm that agent findings are landing in the right session (useful when multiple sessions run in parallel).
When resuming a stalled session after a break.

When NOT to use

For a list of all sessions — use discover list.
To start a new one — use discover start.

Usage

forgeplan discover show <SESSION_ID> [OPTIONS]

Arguments

  <SESSION_ID>   Discovery session identifier

Options

  -h, --help     Print help
  -V, --version  Print version

Examples

forgeplan discover show disc-001
forgeplan discover show disc-001 | less

# Typical mid-run check
forgeplan discover start
# ...agent runs for a while...
forgeplan discover show disc-001

What you see

Typical output blocks:

Session header — ID, created_at, status (active / completed), owner/agent.
Coverage — which source tiers have been touched (code / git / tests / docs), with counts.
Findings by category — decisions, invariants, drift, debt, risks.
Recent findings — latest N entries with summary text.
Next action hint — either “keep scanning” or “ready to complete”.

How it fits

discover show is the read side of the discovery loop. It reads directly from the session store (LanceDB), so it always reflects the latest discover_finding MCP calls without needing a refresh.

discover start → [agent appends via MCP] → discover show → discover complete