trueflow
Trueflow
Use this skill only as an orchestrator for these downstream skills:
trueflow_initializertrueflow_adversarytrueflow_referee
Do not perform substantive evaluation, analysis, judgment, or fixing yourself.
Workspace
- Use
.context/trueflow/as the durable workspace. - Stage outputs:
.context/trueflow/initializer.md.context/trueflow/adversary.md.context/trueflow/referee.md
- Consolidated output:
.context/trueflow/findings.md
Inputs
- original artifacts
- optional domain context
- optional evaluation goal
- optional constraints
If the original artifacts are missing, ask for them before starting.
Handoff Rules
- Preserve the original artifact order across all stages.
- Use
.context/trueflow/initializer.mdas the canonical row scaffold. - Preserve every
Index,Context / Topic, andFINDING-###identifier exactly as emitted bytrueflow_initializer. - Pass prior-stage outputs verbatim to the next stage.
- Do not rewrite, renumber, merge, split, reorder, reinterpret, summarize, or filter any row between stages.
- Do not add commentary outside the required stage inputs and final output.
Workflow
Stage 1
Invoke trueflow_initializer with:
- original artifacts
- optional domain context
- optional evaluation goal
- optional constraints
Require the downstream response to be exactly:
.context/trueflow/initializer.md
Then read the full contents of .context/trueflow/initializer.md.
Stage 2
Invoke trueflow_adversary with:
- original artifacts
- the full contents of
.context/trueflow/initializer.md - any optional domain context still relevant
- any optional evaluation goal still relevant
- any optional constraints still relevant
Require the downstream response to be exactly:
.context/trueflow/adversary.md
Then read the full contents of .context/trueflow/adversary.md.
Stage 3
Invoke trueflow_referee with:
- original artifacts
- the full contents of
.context/trueflow/initializer.md - the full contents of
.context/trueflow/adversary.md - any optional domain context still relevant
- any optional evaluation goal still relevant
- any optional constraints still relevant
Require the downstream response to be exactly:
.context/trueflow/referee.md
Then read the full contents of .context/trueflow/referee.md.
Consolidation
After all three stages succeed, write .context/trueflow/findings.md as a
Markdown table with this exact header:
| Index | Context / Topic | Initializer Finding | Adversary Finding | Referee Verdict |
|---|---|---|---|---|
Consolidation rules:
- Use the rows from
.context/trueflow/initializer.mdas the authoritative row order. - Copy
Index,Context / Topic, andInitializer Findingfrom.context/trueflow/initializer.md. - Copy
Adversary Findingfrom.context/trueflow/adversary.md. - Copy
Referee Verdictfrom.context/trueflow/referee.md. - Do not rewrite any copied cell text.
- Fail if any row count, row order,
Index,Context / Topic, orFINDING-###identifier is inconsistent across files.
Downstream Failure Handling
If any downstream skill is unavailable, fails during execution, or returns malformed output, stop immediately.
Treat output as malformed when any of the following occurs:
- the downstream response is not the exact expected path
- the expected file is missing
- the file does not contain the required table header
- row counts differ across stage files
- any row changes
Index,Context / Topic, orFINDING-###unexpectedly - a stage file fills columns other than its own stage column
Do not continue to later stages after a failure. Do not repair, replace, simulate, approximate, or complete a failed stage.
Return only this exact error format:
Error: <stage_name> failed due to <unavailability|malformed output|execution failure>.
Use stage_name values trueflow_initializer, trueflow_adversary,
trueflow_referee, or trueflow_consolidation.
Final Output
On success, return only the full contents of .context/trueflow/findings.md
exactly as written.
Hard Prohibitions
- Never perform substantive evaluation work.
- Never add findings.
- Never remove findings.
- Never challenge findings.
- Never adjudicate findings.
- Never propose fixes.
- Never substitute your own reasoning for any downstream skill.
- Never rewrite stage cell contents during consolidation.
- Never add any text before or after the final consolidated table.
More from henryqw/skills
gh-autopilot
Run a standalone autonomous GitHub Copilot pull request review loop with explicit stage entry and event logs. Use when Codex should start from a user-selected stage (create PR, monitor review, or address existing comments), execute deterministic cycle transitions, and continue looping until Copilot reports no comments or the configured Stage 2 max-wait limit is reached.
10gh-pr-creation
Create a new GitHub pull request end-to-end when the user asks to open or create a PR. Use when Codex must turn local uncommitted work into a reviewable PR by making multiple scoped commits, running and passing all repository quality gates, renaming the branch so it reflects the changes, creating a Conventional Commits PR title, writing a PR description with summary/rationale/migration steps, and assigning Copilot as reviewer.
6gh-address-copilot-review
Handle GitHub PR review comments when comments are provided by the user as context. Use when Codex must evaluate comments one by one, classify each as actionable or non-actionable or needs clarification, implement only necessary fixes, keep changes scoped per comment, run validation, avoid intermediate pushes, perform one final push for the full batch, resolve addressed threads, respond to rejected comments with rationale, and re-request Copilot reviewer exactly once at the end via gh-assign-copilot-reviewer.
5triangulate
Evaluate supplied artifacts and return a consolidated findings table with evidence-based conclusions. Use this skill when the user wants a proposal, plan, code change, document, prompt, transcript, or other material reviewed through a structured multi-perspective evaluation instead of a single opinion.
3codex-subagent
Dispatch one or more tasks to Codex CLI subagents to save Claude Code tokens. Accepts explicit task descriptions, auto-selects sandbox (read-only vs workspace-write) and reasoning effort (high vs xhigh) based on task type, and collects structured results with durable artifacts.
2gh-pilot
Iteratively drive a PR through GitHub Copilot review using a simple loop with direct `gh` commands and no helper scripts. Reuse existing Copilot feedback first, fetch unresolved thread state via GraphQL, request/re-request Copilot when needed, and require a fresh Copilot pass after pushed fixes.
1