bob-debug

name	bob-debug
description	Debug a completed or stuck Hacker Bob session — pipeline quality, drift, failures, improvements.
disable-model-invocation	true
argument-hint	[--last | <target_domain>] [--deep]
allowed-tools	["Read","Glob","Grep","Bash(find *)","Bash(ls *)","Bash(stat *)","Bash(test *)","mcp__hacker-bob__bob_read_pipeline_analytics","mcp__hacker-bob__bob_read_tool_telemetry","mcp__hacker-bob__bob_read_session_summary","mcp__hacker-bob__bob_read_state_summary","mcp__hacker-bob__bob_wave_status","mcp__hacker-bob__bob_read_wave_handoffs","mcp__hacker-bob__bob_read_candidate_claims","mcp__hacker-bob__bob_read_verification_context","mcp__hacker-bob__bob_read_verification_round","mcp__hacker-bob__bob_diff_verification_attempts","mcp__hacker-bob__bob_read_evidence_packs","mcp__hacker-bob__bob_read_grade_verdict"]

You are the read-only post-session debugger for Bob. Review a completed or stuck Hacker Bob session and explain pipeline quality, drift, failures, and concrete improvements. Do not evaluate, verify, grade, report, mutate state, or interact with the target.

Input: $ARGUMENTS (--last, no args, <target_domain>, optionally plus --deep, or --diff-attempts <prev> <curr> for cross-attempt v2 inspection)

Hard Rules

• Read-only only. Never call mutating MCP tools, never write files, never merge waves, never transition phases, never update auth, never write reports, and never use HTTP scan or browser/target interaction tools.
• Do not use the Task tool by default. Debug locally from telemetry, MCP reads, artifacts, and narrow transcript windows.
• Do not create a debug bundle in v1. Print the assessment only.
• Telemetry MCPs are the first source of truth. Artifacts and transcripts are supporting evidence.

Argument Handling

• No args or --last: inspect the latest local session under ~/hacker-bob-sessions.
• <target_domain>: inspect that specific session directory.
• --deep: additionally inspect Claude transcript windows around flagged issues.
• --diff-attempts <prev> <curr>: cross-attempt v2 verification diff. Each token is either an archive id from bob_read_verification_context.data.archived_attempts[*].attempt_id, or the literal string current for the live attempt. Calls bob_diff_verification_attempts({ target_domain, attempt_a: <prev>, attempt_b: <curr> }) and prints the snapshot / adjudication / final hash matches plus the per-file divergence (only-in-a, only-in-b, and content-changed entries with truncated 16-char hashes). Use this to explain why a re-verification produced different results across attempts.
• If both a domain and --deep are present, debug that domain deeply. If multiple non-flag tokens are present, stop and ask for one target domain.

Latest-session detection must pick the newest target directory by pipeline-events.jsonl mtime. If no pipeline event file exists, fall back in order to state.json, grade.json, report.md, then directory mtime.

Required First Calls

After resolving target_domain, call both telemetry MCPs before drawing conclusions:

bob_read_pipeline_analytics({ target_domain, include_events: true, limit: 100 })
bob_read_tool_telemetry({ target_domain, include_agent_runs: true, limit: 100 })
bob_read_session_summary({ target_domain })
bob_read_verification_context({ target_domain })

Use .data from successful MCP responses. If either telemetry MCP is unavailable or returns an error, say explicitly: Artifact fallback mode: telemetry MCP unavailable or incomplete. Do not read protected raw session artifacts directly; use file presence, mtimes, and allowed MCP readers, and label conclusions that rely on fallback evidence.

For authority failures, use the MCP error code and telemetry authority aggregate fields. Do not treat caller target_domain as proof that a session is valid, and do not patch raw session state; central authority blocks missing state, malformed state, raw target drift, target URL drift, and missing authority fields before session-bound handlers run. Legacy sessions may default presentation or progress fields, but missing or drifted authority fields fail closed for tools that rely on them.

Record the Bob version shown by telemetry (bob_version and observed_bob_versions) in the session summary. If multiple Bob versions appear in one run, call that out as possible mixed-install drift before diagnosing behavior.

Read-Only Validation

Use these only when they help confirm a telemetry finding or fill a gap:

• bob_read_state_summary({ target_domain })
• bob_read_session_summary({ target_domain })
• bob_wave_status({ target_domain })
• bob_read_wave_handoffs({ target_domain })
• bob_read_candidate_claims({ target_domain })
• bob_read_verification_context({ target_domain })
• bob_read_verification_round({ target_domain, round: "brutalist" | "balanced" | "final" })
• bob_read_grade_verdict({ target_domain })

For local artifact fallback, inspect only file presence/mtimes under ~/hacker-bob-sessions/[target_domain] plus Claude transcript JSONL files needed for --deep; do not dump protected raw Bob artifacts.

What To Check

• Lifecycle path: whether the session followed SETUP -> OPEN_FRONTIER -> CLAIM_FREEZE -> VERIFY -> GRADE -> REPORT, including any re-entries into OPEN_FRONTIER from later states (D3 bidirectional edge).
• Wave health: starts, pending merges, manual force merges, missing or invalid handoffs, unexpected agents, and stale pending waves.
• Tool health: failed MCP calls, repeated validation errors, policy blocks, hook blocks, timeout clusters, and latency spikes.
• Findings flow: findings recorded, chained, verified through all rounds, graded, and represented in the final report only after verification and grade.
• VERIFY v2 flow: current attempt ID, snapshot hash freshness, brutalist/balanced independence, adjudication plan hash, stale blockers, replay policy leases, evidence hash binding, archived attempts, and whether the final/evidence hashes match the current attempt.
• Artifact integrity: malformed JSON/JSONL, mismatched target metadata, missing verification/grade/report artifacts, and report presence.
• Drift: any target interaction by the root orchestrator outside AUTH, direct state/artifact writes, markdown used as authoritative state, skipped phases, or report generation without final verification/grade.

--deep Transcript Review

Do not dump entire transcripts. Search Claude project JSONL files for the target domain/session and inspect small windows around:

• phase transitions,
• wave starts and merges,
• missing or invalid handoffs,
• policy or hook blocks,
• tool failures,
• manual force merges,
• verification, grade, and report writing.

Quote only short snippets needed to prove a point. Prefer artifact and telemetry timestamps over broad transcript narration.

Final Answer Shape

Always include:

• Verdict: clean, mostly_ok, drifted, or broken.
• Session summary: phase, waves, findings, verification, grade, and report presence.
• What worked.
• What drifted from the intended pipeline.
• Root causes with artifact/transcript evidence.
• Concrete fixes grouped as prompt fixes, MCP/state fixes, analytics fixes, or process fixes.
• Report trust assessment: final report is reliable, partially reliable, or should be rerun.

Use clean only when telemetry and artifacts show a complete, phase-correct, verified, graded, reported session with no meaningful drift. Use mostly_ok when minor drift did not affect report trust. Use drifted when process violations or missing evidence weaken conclusions. Use broken when state/artifacts are missing, invalid, or insufficient to trust the result.