handoff

A command that creates an end-of-phase handoff document from a template and updates the project's build status. It also asks the team to record unresolved decisions.

In plain words
What is it for?
Use it at the end of a named project phase to run pre-flight checks, warn about missing gates or bypassed design approval, generate the handoff, and prompt decision tracking.
Why use it?
It collects what was completed, checks required approval and deployment markers, and makes open issues visible before work moves to the next phase.

Command

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/rajconnects/founder-stack/handoff
Clone the repo
git clone --depth 1 https://github.com/rajconnects/founder-stack
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,313 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00024 $0.01313
Opus 5 $0.00012 $0.00656
Sonnet 5 $0.00005 $0.00263
Haiku 4.5 $0.00002 $0.00131

Measured yesterday against content hash 4bac0080206c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

handoff scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

workflow/commands/handoff.md · 62 lines

How it starts

The opening of the file, as written. The whole thing — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are running the end-of-phase handoff.

Arguments: $ARGUMENTS

Steps

  1. If $ARGUMENTS is empty, fail fast: "handoff requires a phase identifier. Example: /handoff 6a."

  2. Read .claude/project.json. Extract handoff_template, handoff_output_dir, build_status_file, decision_traces.

  3. Pre-flight gate check. Look for session markers:

    • .claude/.design-gate-passed (optional but expected)
    • .claude/.schema-gate-passed (only if this phase touched migrations)
    • .claude/.deploy-gate-passed-staging (required if phase is shipping a hosted service)
    • .claude/.publish-gate-passed-<artifact> (required if phase is shipping a packaged artifact, one per entry in release_artifacts)

    If any required marker is missing, warn the user and ask whether to proceed anyway. Do NOT silently generate a handoff doc for an unverified phase.

3b. Design-bypass scan. Glob .claude/.design-bypass-*. For each marker found, extract the scope slug from the filename. Print loudly: "⚠ Design approval bypassed for: . The operator invoked /frontend-build --bypass for these scopes. Confirm justification before completing handoff." Ask the user for a one-line justification per bypass; record under "Known risks / deferred items" in the generated handoff doc (step 5). After the handoff doc is written successfully (step 5), delete the bypass markers — they should not carry into the next phase.

3a. Schema-of-record drift check (if schemas_of_record is configured). Protocol/schema bugs are easy to miss because the truth lives in multiple places that drift independently. For each entry in schemas_of_record:

  • Run git diff --name-only <phase-start-ref> HEAD to see what changed in this phase.
  • If canonical is in the diff, ALL shadowed_by paths SHOULD also be in the diff. Surface mismatches: "You changed <canonical> but not <shadowed_by[i]>. Producer/consumer drift risk — confirm intentional."
  • If only some shadowed_by paths changed but not the canonical, surface the inverse: "You changed <shadowed_by[i]> but not the canonical <canonical>. The schema spec may now be out of sync with implementation." This is a soft check — the user confirms intent. Some changes legitimately affect only one site (e.g., a runtime validator becoming more lenient than the spec on purpose). Surface the discrepancy; let the user decide.
  1. Read the template at handoff_template. Read the current build_status_file to understand what's been shipped.

  2. Generate handoff doc. Output path: <handoff_output_dir>/phase-<ARGS>-handoff.md. Fill in the template with:

    • Phase name, start date (ask user if unknown), end date (today per CLAUDE.md currentDate).
    • What shipped (grep git log --oneline --since=<phase start> for commits in phase).
    • Gate results (from session markers + user recall).
    • Known risks / deferred items (ask user; do not fabricate).
    • Open decisions surfaced during the phase (see step 6).
    • Next-phase entry criteria.
  3. Open decisions scan. Ask the user: "Did this phase surface any decisions that need capturing? (yes/no/list)". If yes or a list is given, capture each one as a trace file in decision_traces (path from project.json) using the schema documented in docs/decision-traces.mdid, date, title, status, decided_by, decision, alternatives_rejected[], rationale, revisit_triggers[]. If the project has a decision-trace-capture skill installed, delegate to it instead. One file per decision.

  4. Update build-status. Append or update the section for this phase in build_status_file: items shipped, items deferred, known risks. Preserve existing structure.

  5. Close the session claim. Read .claude/coordination.json:

    • Find the row where phase matches $ARGUMENTS and status: active. If multiple match (rare), prefer the one whose worktree path equals $PWD or is the current worktree per git rev-parse --show-toplevel.
    • If found: set status: completed, set completed to now (ISO-8601), preserve all other fields.
    • If the row's worktree path is set and exists, ask the user: "Remove worktree at <path>? (yes/no)". On yes, run git worktree remove '<path>' from the main repo root. Do not force-remove if uncommitted changes exist — surface and stop.
    • If no matching row: warn but continue (this handoff may pre-date the start-build flow).

Read the full file on GitHub · 62 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 62 lines · 24 tokens per session scan A 4bac0080206c

Subscribe to this mod's changes

handoff is a command published in the GitHub repository rajconnects/founder-stack (2 stars, last pushed 1mo ago), licensed MIT. It adds 24 tokens to every session and 1,313 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.