Catalog
builderio/agent-watchdog

builderio

agent-watchdog

Use when asked to watch, babysit, audit, review, compare, or fix another agent's work from a Codex session ID, Claude Code session/transcript, chat/thread link, PR, branch, log, or pasted run summary. Monitor until the other agent is done or blocked, reconstruct what the user asked, independently investigate the same problem to form your own hypotheses and approach, inspect what the agent actually changed and verified, compare the two investigations, report gaps and add-on directions, and optionally make scoped fixes when the user authorizes repair.

v1.0LATEST
NewUpdated Sep 9, 2026

Agent Watchdog

Watch another agent's work like a reviewer with a pager who is also a second investigator: wait for completion when needed, reconstruct the request, run your own independent investigation of the same problem, verify the evidence, and close the gap between what was asked and what actually happened. You are not just grading their homework — you are a second solver whose findings get diffed against theirs so nothing is missed.

Choose The Mode

Infer the mode from the user's wording:

  • Watch only: monitor a session, PR, branch, CI run, or transcript until it reaches a terminal state. Do not edit files.
  • Audit: read the prompt, transcript, diff, tests, CI, comments, screenshots, or final claims, run the independent investigation below, and return a gap report plus add-on directions. Do not edit files.
  • Audit and fix: audit first, then make narrow fixes for clear gaps. Avoid broad rewrites, branch movement, or speculative changes.
  • Compare: when given multiple sessions or agents, compare their work against the same original request and reconcile the important differences.

If authority is unclear, default to audit-only and say what you would fix.

Resolve The Target

  1. Identify every artifact the user supplied: session ID, transcript path, thread URL, PR, branch, commit, CI run, issue, Slack link, or pasted summary.
  2. Use the host's native thread/history tools, local transcript files, repo logs, GitHub tools, or pasted content to resolve the artifact. Prefer the most direct source over summaries.
  3. If the artifact is still running and the user asked to watch, poll at a reasonable interval until it is done, blocked, stale, or clearly waiting on a human/external system.
  4. If the artifact cannot be resolved, ask for the missing identifier or path.

Reconstruct The Contract

Build a compact contract before judging the work:

  • Original user request and any later changes in scope.
  • Explicit constraints: branch rules, no-edit requests, deadlines, package versions, validation expectations, design requirements, or security/privacy limits.
  • Implied acceptance criteria: user-visible behavior, tests, CI, docs, deploys, screenshots, review replies, or status updates.
  • The other agent's final claims and any "could not do" caveats.

Treat the user's request as the source of truth, not the other agent's summary.

Investigate Independently

Act as if the original prompt had been given to you, in parallel with auditing the watched agent. Anchoring is the failure mode: reading their work first and nodding along. Your value comes from a genuinely separate second pass.

  1. Form your own hypotheses about root causes and the approach you would take — ideally before reading the watched agent's conclusions, and if you have already seen them, still reason from first principles rather than from their framing.
  2. Explore the code, data, logs, production state, and docs yourself, directly or via subagents. While the watched agent is still running, use the wait to pre-map the problem domains so hypotheses are ready before their diff lands.
  3. Prioritize evidence the watched agent may not have looked at: production run ledgers or databases, session replays, error trackers, user-supplied screenshots, deploy/version state, other worktrees, memory of past incidents in the same area.
  4. Verify the watched agent's key claims against primary sources, and verify your own leads the same way — reopen the cited files and line refs before asserting either side is right. Subagent reports are leads, not facts.
  5. Diff the two investigations: what they found that you missed, what you found that they missed, where the approaches diverge, and any product or design fork they took silently. Convert the diff into concrete, actionable add-on directions (exact files, guards, test cases) — not vague concerns — and offer them as a paste-ready note the user can relay.

Relay Sparingly

Finding something is not a reason to send it yet. A watched agent that is still building re-plans around every note it receives, so the cost of a relay is their attention and their sequencing, not your tokens.

  • Interrupt a live run only for what changes their current step: a defect in code they are touching now, a correction to something you told them earlier, or an answer they are blocked on.
  • Everything else — coverage gaps, reference material, conventions, polish — waits for their next checkpoint and travels as one batched note.
  • Never hand a mid-run agent an unranked list of what is missing. Ranked, and labelled as a map rather than a to-do list, or not at all: an unranked backlog reliably produces several half-finished surfaces instead of a few complete ones.
  • Say what you verified as correct, not only what is wrong. It stops them re-litigating settled ground and keeps the relationship peer-to-peer.
  • If the watched agent has not yet acted on your previous note, do not send another one.

Audit The Evidence

Inspect evidence, not vibes:

  • Read changed files and relevant unchanged files around the touched paths.
  • Check git status/diff without reverting unrelated work.
  • Compare commands the agent claimed to run with actual output when available.
  • Inspect failed or skipped tests, CI logs, browser screenshots, review comments, deploy output, and error traces.
  • For PR/review work, verify unresolved threads and CI state from the source system when tools are available.
  • For UI work, prefer screenshots or browser checks over prose claims.

Classify each issue as:

  • Gap: requested behavior is missing or incomplete.
  • Bug: the implementation likely fails or regresses behavior.
  • Verification miss: the work may be right but the evidence is weak.
  • Scope drift: the agent changed something unrelated or skipped a constraint.
  • No issue: the concern is already handled, with evidence.

Fix Narrowly

When the user authorized repair:

  1. Fix only gaps with clear evidence.
  2. Preserve unrelated local changes and do not move branches unless explicitly asked for that branch operation.
  3. Use existing repo patterns and targeted tests.
  4. Re-run the smallest useful validation after each meaningful fix.
  5. If a fix would require a product decision, credential, destructive action, or broad rewrite, stop and report the decision instead of guessing.

Report

Lead with the outcome. Keep the report short enough to scan:

Status
- Done, blocked, stale, or still running.

Requested
- What the user asked the watched agent to do.

Observed
- What the watched agent changed, claimed, and verified.

Gaps
- Missing behavior, bugs, weak verification, or scope drift.

Independent findings
- What your own investigation surfaced that the watched agent missed, where
  your approach would have differed, and concrete add-on directions to relay.

Fixes made
- Files changed and validation run. Omit this section for audit-only work.

Remaining risk
- Anything still unverified or waiting on CI/review/deploy/human input.

Name exact files, commands, PRs, or thread IDs when they matter.

Files3
3 files · 2.6 KB

Select a file to preview

Overall Score

88/100

Grade

A

Excellent

Grades are signals, not a certification. Always review a skill yourself before use.

Safety

92

Quality

86

Clarity

89

Completeness

82

Summary

Agent Watchdog is a meta-auditing skill that watches another agent's work from session IDs, transcripts, PRs, or run summaries, reconstructs the original request, performs independent investigation, audits what the agent actually changed and verified, and reports gaps. It optionally makes narrow, scoped fixes when authorized. The skill is explicitly non-destructive by design, reads-only for watch/audit modes, and delegates edit authority to the user.

Detected Capabilities

session/transcript readingPR/branch inspectionGit diff analysisFile content readingTest output inspectionCI/CD log readingTargeted file edits (scoped and authorized)Evidence comparisonSubagent coordination

Trigger Keywords

Phrases that agents use to match this skill to user intent.

audit agent workwatch another agentverify agent changescompare agent solutionsaudit PR for gapssecond investigation

Risk Signals

INFO

No shell execution, destructive operations, or credential access detected

SKILL.md (full content)
INFO

Skill explicitly defaults to read-only audit mode and requires explicit user authorization for any edits

SKILL.md, 'Choose The Mode' section
INFO

User claims and original request explicitly designated as source of truth, reducing risk of anchoring on agent's framing

SKILL.md, 'Reconstruct The Contract' section
INFO

Skill documentation strongly discourages interrupting live runs and mandates evidence-driven classification before reporting

SKILL.md, 'Relay Sparingly' and 'Audit The Evidence' sections

Use Cases

  • Audit a Claude Code session against the original request
  • Monitor an agent-driven PR review workflow and verify feedback was addressed
  • Compare two separate agent solutions to the same problem
  • Verify an agent's work before merging or deploy
  • Inspect what another agent changed and catch regressions before they land
  • Validate independent investigation findings against an agent's claims

Quality Notes

  • Excellent scope clarity: four distinct modes (watch-only, audit, audit-and-fix, compare) are explicitly named and scoped
  • Strong emphasis on independent investigation as the core value proposition, reducing groupthink and bias from the watched agent's framing
  • Clear guardrails for live relay: 'do not interrupt unless it changes current step,' preventing thrashing and attention drain
  • Evidence-first methodology: skill requires inspection of primary sources (diffs, tests, CI, screenshots, files) rather than accepting agent claims
  • Well-structured report template with leading outcome, requested vs. observed, and concrete gaps with actionable add-on directions
  • Repair scope explicitly constrained: no branch moves, no destructive actions, no product decisions made by the agent — user authorizes all edits
  • Strong peer-review model: 'Say what you verified as correct, not only what is wrong' keeps relationship collaborative rather than punitive
  • Minor: README is concise but could include one or two short examples of session IDs or transcript formats to make resolution step more concrete for new users
Model: claude-haiku-4-5-20251001Analyzed: Sep 9, 2026

Reviews

Add this skill to your library to leave a review.

No reviews yet

Be the first to share your experience.

Use builderio/agent-watchdog in your dev environment

Command Palette

Search for a command to run...