TRIP Upgrade logo

TRIP Upgrade

Community
PiLastDigit
TRIP-upgrade

Upgrade TRIP workflow skills to a newer version while preserving project customizations

Overview

PublisherPiLastDigit
RepositoryTRIP-workflow
Skill nameTRIP-upgrade
Stars
636
Forks
91
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by PiLastDigit on GitHub. Read the source before you install it.

Installation

Install the TRIP Upgrade AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/PiLastDigit/TRIP-workflow/tree/master/skills/TRIP-upgrade
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/PiLastDigit/TRIP-workflow.git /tmp/TRIP-workflow
mkdir -p .claude/skills
cp -r /tmp/TRIP-workflow/skills/TRIP-upgrade .claude/skills/TRIP-upgrade
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable TRIP Upgrade in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads TRIP Upgrade on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so TRIP Upgrade is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

TRIP Upgrade Mode

You are now in upgrade mode — merging a newer version of the TRIP workflow into this project's existing, customized TRIP skills.

The Problem

Each project's TRIP skills have two interleaved layers:

  1. Workflow skeleton — steps, Codex integration, file structure, process flow
  2. Project customizations — test commands, checklist sections, version file, technical considerations, guidance sections

A naive copy would destroy layer 2. This skill recovers the generic template the project was installed from, computes the project's customizations as a diff against it, and replays that diff onto the new template with a git three-way merge — one command per file, exact, and verifiable. Manual extraction of the customizations remains as the path for pre-2.8.1 installs, which carry no version stamp.

Prerequisites

Nothing to copy. The skill fetches the TRIP-workflow repository itself and upgrades to its latest release by default.

$ARGUMENTS, optional:

  • a version (2.9.0 or v2.9.0) pins the target release instead of the latest;
  • a path to a local skills/ folder (a fork, an offline copy) is used as the staged source instead of fetching. In that case the base template (Phase 2) still needs the clone; without network, use the Fallback path.

Phase 1: Inventory

1.1 Fetch the Latest Release

One clone serves the whole upgrade — it provides the new skills now and the base template in Phase 2:

bash
TRIP_SRC=$(mktemp -d)
git clone -q --no-checkout https://github.com/PiLastDigit/TRIP-workflow "$TRIP_SRC"
TARGET=$(git -C "$TRIP_SRC" tag --list 'v*' --sort=-v:refname | head -1)   # or v$ARGUMENTS when a version was given
S=$(mktemp -d); git -C "$TRIP_SRC" archive "$TARGET" skills | tar -x -C "$S" --strip-components=1

$S is the staged source for the rest of this skill. If a path was passed instead, S=<that path> and skip the archive. If the clone fails (offline, no access), tell the user and stop — or take the Fallback path with a local folder.

1.2 Detect Versions

Since v2.8.1 every SKILL.md carries metadata.trip-version in its frontmatter. Read both sides:

bash
# Installed (one line per distinct version; more than one line = a mixed install)
grep -h "^  trip-version" .claude/skills/*/SKILL.md | sort | uniq -c

# Staged
grep -h "^  trip-version" "$S"/*/SKILL.md | sort -u
  • Field present on both sides: report installed X → target Y. If X equals Y, tell the user the project is already on the latest release and stop; remove $TRIP_SRC and $S.
  • Field missing in the installed copy: the install predates v2.8.1. Describe it from structural hints — --speedrun in TRIP-1-plan = 2.7.x, checklist.md + TRIP-3-release present = v2, codex-* skills absent = v1 — and plan on the Fallback: Manual Extraction path, since no base template can be fetched.
  • Mixed versions: list the outliers; they are usually skills skipped in a previous upgrade and should be merged like any other.

1.3 Categorize Skills

List all skill folders in both locations:

bash
# Currently installed
ls -d .claude/skills/*/

# New (staged)
ls -d "$S"/*/

Categorize each skill into one of:

CategoryMeaningAction
NewExists in $S onlyCopy directly
RemovedExists in installed onlyWarn user, leave in place
UnchangedIdentical in bothSkip
Updated — pure workflowChanged, but no project customizationsReplace directly
Updated — customizedChanged, AND contains project-specific contentExtract → merge → replace

Pure workflow skills (no project customizations): TRIP-compact, TRIP-hotfix, TRIP-research, TRIP-init, codex-implement, codex-plan-review, codex-code-review

Exception — model defaults: codex-plan-review/scripts/_common.sh holds the per-flow Codex model/effort/service-tier defaults, which the user may have tuned. Before replacing, diff the installed _common.sh against staging — if the model/effort/tier values differ from the generic defaults, carry the user's values into the new file.

Customized skills (have project-specific content): TRIP-1-plan, TRIP-2-implement, TRIP-3-release, TRIP-review, TRIP-test

Renamed in TRIP v2 — when the installed folder uses an old name, treat it as the same skill under its new name (merge into the new name, then delete the old folder):

Installed (old)Target (new)
TRIP-3-reviewTRIP-review
TRIP-4-testTRIP-test

TRIP-3-release is new in v2 but its project values (version file, week anchor, tutorial config) are extracted from the old TRIP-2-implement's post-implementation steps — categorize it as customized even though no folder exists yet.

Other non-TRIP skills in $S (e.g. future additions): Treat as new, or as pure workflow if they already exist.

For each skill, diff the installed vs new version to confirm whether it actually changed:

bash
diff -rq .claude/skills/<skill>/ "$S"/<skill>/

1.4 Present Inventory

Show the version line, then a summary table to the user:

TRIP 2.7.4 → 2.8.1 (installed version read from metadata.trip-version)

Skill                 | Status              | Action
--------------------- | ------------------- | ------
TRIP-1-plan           | Updated (customized) | Extract + merge
TRIP-2-implement      | Updated (customized) | Extract + merge
TRIP-3-release        | New (customized)     | New template + values from old TRIP-2
TRIP-review           | Renamed + updated    | Extract + merge, delete TRIP-3-review/
TRIP-test             | Renamed + updated    | Extract + merge, delete TRIP-4-test/
TRIP-compact          | Unchanged            | Skip
TRIP-hotfix           | Unchanged            | Skip
TRIP-init             | Updated (pure)       | Replace
TRIP-research         | Unchanged            | Skip
codex-plan-review     | New                  | Copy
codex-code-review     | New                  | Copy
codex-implement       | New                  | Copy

AskUserQuestion: "Here's the upgrade plan. Proceed?" Options: "Yes, start upgrade" (recommended) / "Let me review the new files first" / "Abort"


Phase 2: Obtain the Base Template

The merge needs the generic skills of the version currently installed (the "base"). The project never holds them — only the customized copy — so fetch them from the TRIP-workflow repository.

2.1 Reuse the Clone

$TRIP_SRC from Phase 1.1 already holds every release tag:

bash
git -C "$TRIP_SRC" tag --sort=-v:refname | head -20

2.2 Identify the Base Tag

The base tag is v<installed trip-version> from Phase 1.2. No stamp means a pre-2.8.1 install: the base cannot be named with certainty, so skip to Fallback: Manual Extraction below. This happens once per project — every install from 2.8.1 on carries the stamp.

bash
BASE=$(mktemp -d)
git -C "$TRIP_SRC" archive "v$INSTALLED_VERSION" skills | tar -x -C "$BASE" --strip-components=1

2.3 Report

"Base template v2.8.2 fetched. Upgrading v2.8.2 → v2.9.0."


Phase 3: Three-Way Merge

For every file in $S, state folders excluded, decide by presence:

InstalledBaseAction
yesyesgit merge-file -p installed base staged → merged file
yesnofile is new upstream but the project also has one: merge with an empty base (git merge-file -p installed /dev/null staged)
no—copy staged
(installed only)—leave in place, list as "Removed upstream"

Build everything in a scratch directory first; nothing under .claude/skills/ changes until Phase 4 passes.

bash
MERGED=$(mktemp -d); I=.claude/skills   # $S = staged source from Phase 1.1
( cd "$S" && find . -type f -not -path '*/state/*' ) | sed 's|^\./||' | while read -r f; do
  mkdir -p "$MERGED/$(dirname "$f")"
  if [ -f "$I/$f" ]; then
    b="$BASE/$f"; [ -f "$b" ] || b=/dev/null
    git merge-file -p -L installed -L base -L staged "$I/$f" "$b" "$S/$f" > "$MERGED/$f" || echo "CONFLICT $f"
  else
    cp "$S/$f" "$MERGED/$f"
  fi
done
grep -rl '^<<<<<<< ' "$MERGED" || echo "no conflicts"

Conflicts happen only where a workflow change and a customization touch the same lines. Open the marked region, keep the new workflow text and re-apply the project's intent inside it, remove the markers. Never resolve by taking one side wholesale.

Renamed skills (TRIP-3-review → TRIP-review, TRIP-4-test → TRIP-test): treat the old folder as the installed side of the new name, then delete the old folder in Phase 5. TRIP-3-release on a v1 install: no installed side exists; copy staged and fill its placeholders from the old TRIP-2-implement post-implementation steps (see the Fallback section for what to look for).


Phase 4: Review and Validate

4.1 Prove the Merge

Two diffs per merged file, both must be short and readable:

bash
# what the project customized — must contain ONLY project-specific content
diff "$S/$f" "$MERGED/$f"
# what the upgrade changes — must contain ONLY workflow changes (and the new frontmatter)
diff "$I/$f" "$MERGED/$f"

Read both for every customized skill (TRIP-1-plan, TRIP-2-implement, TRIP-3-release, TRIP-review/*, TRIP-test) and for any pure skill whose first diff is not empty — a project sometimes tweaks a "pure" skill by a line, and the merge carries that over. A customization appearing in the second diff as a removal, or workflow text appearing in the first diff, means the merge went wrong for that file — fix it before continuing.

4.2 Placeholder Check

Scan the merged files for leftover placeholders:

bash
grep -rn --exclude-dir=TRIP-init --exclude-dir=TRIP-upgrade '\[ADAPT_TO_PROJECT\|\[PROJECT_NAME\]\|\[VERSION_FILE\]\|\[WEEK_ANCHOR_DATE\]\|\[TEST_COMMAND\]\|\[LINT_COMMAND\]\|\[TYPECHECK_COMMAND\]\|\[TUTORIAL_STEP\]\|\[MAIN_BRANCH\]' "$MERGED"/TRIP-*/

TRIP-init and TRIP-upgrade mention every placeholder by name in their own instructions, so they are excluded.

If any are found, fill them from context or ask the user.

4.3 Cross-Reference Check

  • checklist.md section names must match cr-template.md checklist section names
  • codex-code-review/prompts/start.tpl and resume.tpl reference .claude/skills/TRIP-review/checklist.md — confirm it exists, and that no template still points at the old TRIP-3-review/ path
  • codex-code-review/prompts/synthesize.tpl and codex-code-review/SKILL.md reference .claude/skills/TRIP-review/cr-template.md — confirm it exists
  • TRIP-1-plan references codex-plan-review/scripts/start.sh and resume.sh; TRIP-2-implement and TRIP-research invoke each skill's own wrapper scripts — confirm codex-implement, codex-code-review, and codex-ask each ship scripts/{start,resume,reset,show}.sh (they pin STATE_DIR and delegate to codex-plan-review), and that codex-plan-review/scripts/key.sh exists

4.4 Present Summary

Show what changed:

Upgrade complete:
- New skills added: [list]
- Skills updated: [list]
- Skills unchanged: [list]
- Project customizations preserved: [list key ones]

Run the version-stamp check from Frontmatter and Version Stamp on $MERGED.

AskUserQuestion: "Merge verified. Apply to .claude/skills/?" Options: "Apply" (recommended) / "Show me the diffs first" / "Abort"

If "Show me the diffs first": present the Phase 4.1 diffs. If "Abort": remove the scratch directories, nothing was changed.


Phase 5: Apply and Clean Up

bash
# copy merged files over the install; state/ folders are never touched
rsync -a --exclude 'state/' "$MERGED"/ .claude/skills/

Then delete renamed-away folders (TRIP-3-review/, TRIP-4-test/) if any, remove the scratch directories ($TRIP_SRC, $S unless it is a user-supplied path, $BASE, $MERGED — use trash-put where available, otherwise rm -r), and confirm the stamp once more on the live install:

bash
grep -h "^  trip-version" .claude/skills/*/SKILL.md | sort | uniq -c

Report:

"TRIP workflow upgraded vX → vY. N files merged, M copied, 0 conflicts. git diff .claude/skills/ shows the change before you commit."


Fallback: Manual Extraction

Use this path when the installed skills carry no metadata.trip-version (pre-2.8.1 install). It needs a staged source $S — the fetched release from Phase 1.1, or a local folder passed as argument. It rebuilds each customized skill from the new template by hand.

F.1 Extract Project Context

Before touching any installed files, read every customized skill and extract all project-specific values into a context block. This is your safety net — everything here gets re-injected later.

F.1.1 Read All Installed Skills

Read every file in the installed skills directory that will be affected.

F.1.2 Extract Customizations

Build a context block by extracting these values from the installed skills:

From TRIP-1-plan/SKILL.md:

  • PROJECT_NAME — the text that replaced [PROJECT_NAME] (appears in the # Planning Mode header and **planning mode** for line)
  • TECHNICAL_CONSIDERATIONS — the full content of the ## Technical Considerations section in the plan template (everything between ## Technical Considerations and the next ## heading)
  • GUIDANCE_SECTIONS — everything after the plan template's closing section that replaced [ADAPT_TO_PROJECT: Guidance Sections] (project-specific per-component guidance at the bottom of the file)
  • DOC_IMPACT_CANDIDATES — the candidates list inside the plan template's ## Documentation Impact section, if present (may not exist in pre-v2.5 installs)
  • PLAN_PREREQUISITES — any additional living-docs lines in the Prerequisites list beyond ARCHI.md, if present

From TRIP-2-implement/SKILL.md (in v1 installs, the release values below live in its Post-Implementation steps; in v2 installs they live in TRIP-3-release/SKILL.md):

  • PROJECT_NAME — (confirm matches TRIP-1-plan)
  • VERSION_FILE — the text that replaced [VERSION_FILE] in Step 2
  • WEEK_ANCHOR_DATE — the date that replaced [WEEK_ANCHOR_DATE] in Step 1
  • TUTORIAL_CONFIG — if tutorials are enabled: the full Tutorial step block with user context. If disabled: note "tutorials disabled"
  • LINT_COMMAND — if present (may not exist in older versions)
  • TYPECHECK_COMMAND — if present
  • TEST_COMMAND — if present

From TRIP-review/SKILL.md — or TRIP-3-review/ in v1 installs (checklist.md if already split):

  • REVIEW_CHECKLIST — the full checklist content. In older versions this is inline in SKILL.md. In newer versions it's in checklist.md. Extract it wherever it lives.
  • CR_TEMPLATE — if cr-template.md exists, extract it. Otherwise note "no template file — using inline template"

From TRIP-test/SKILL.md — or TRIP-4-test/ in v1 installs:

  • TEST_COMMANDS — the full Commands section
  • TEST_STRUCTURE — the test structure description
  • TESTING_PRIORITIES — the full testing priorities section
F.1.3 Present Extracted Context

Show the user a summary of what was extracted:

Extracted project context:
- Project name: [name]
- Version file: [path]
- Week anchor: [date]
- Tutorials: [enabled/disabled]
- Test commands: [lint] / [typecheck] / [test]
- Checklist sections: [count] sections ([list names])
- Guidance sections: [count] sections ([list names])
- Technical considerations: [count] items

AskUserQuestion: "Extracted project context looks correct?" Options: "Yes, continue" / "No, let me correct something"

If "No": let the user specify corrections, update the context block.


F.2 Handle Structural Migrations

Before merging, handle any structural changes between the old and new workflow versions. Read both old and new files to detect what changed structurally.

F.2.1 Checklist Extraction (TRIP-review)

Old structure (early v1): Checklist inline in TRIP-3-review/SKILL.md New structure (v2): Checklist in separate TRIP-review/checklist.md, template in TRIP-review/cr-template.md (late-v1 installs have these same files under TRIP-3-review/)

If the installed version has the checklist inline in SKILL.md (no separate checklist.md):

  1. The extracted REVIEW_CHECKLIST from F.1 is the project-customized checklist
  2. It will be injected into the new checklist.md in F.3

If the installed version already has checklist.md:

  1. The extracted content is already in the right format
  2. Merge normally in F.3
F.2.2 Codex Integration (TRIP-1-plan, TRIP-2-implement)

Old structure: No Codex review steps New structure: TRIP-1-plan has Step 3 (Codex plan review), TRIP-2-implement has Codex Code Review section

These are pure workflow additions — no project-specific content to migrate. They will be applied from the new template. The only project-specific part is the test commands in TRIP-2-implement's Codex pre-step, which come from the extracted context.

F.2.3 Codex Skills (codex-plan-review, codex-code-review, codex-implement)

If not installed yet, these are entirely new — copy from $S directly. The review skills reference TRIP-review/checklist.md and TRIP-review/cr-template.md, which will be populated with project content. If already installed (late-v1), replace as pure workflow (see the _common.sh exception in Phase 1.3) — v1 prompt templates point at the old TRIP-3-review/ paths and must be replaced with the v2 versions.


F.3 Customized Skills — Rebuild From Template

For each customized skill, take the new template from $S and inject the extracted project context from Phase 2.

General approach: Read the new template file. Find each placeholder or generic section. Replace with the corresponding extracted value. Write the result.

TRIP-1-plan/SKILL.md
  1. Start from the new template ($S)
  2. Replace [PROJECT_NAME] with extracted PROJECT_NAME
  3. Replace the generic ## Technical Considerations block in the plan template with extracted TECHNICAL_CONSIDERATIONS
  4. Replace the [ADAPT_TO_PROJECT: Guidance Sections] comment block with extracted GUIDANCE_SECTIONS
  5. Resolve the [ADAPT_TO_PROJECT] marker in the ## Documentation Impact section: use extracted DOC_IMPACT_CANDIDATES if present; otherwise build the list from the project's living docs (README, module READMEs, operations/user manuals, reference specs next to the code, contributor guides), one bullet per doc with when it's affected
  6. Resolve the [ADAPT_TO_PROJECT] line in the Prerequisites list: use extracted PLAN_PREREQUISITES if present; otherwise list any living docs a plan must respect beyond ARCHI.md, or delete the line if there are none
TRIP-2-implement/SKILL.md
  1. Start from the new template ($S)
  2. Replace [PROJECT_NAME] with extracted PROJECT_NAME
  3. Replace [LINT_COMMAND], [TYPECHECK_COMMAND], [TEST_COMMAND] in the Testing Gate with extracted commands
    • If the old version didn't have Codex review (no test commands extracted), check the old TRIP-4-test for test commands, or ask the user
  4. Adapt the Integration impact check comment block to the project's integration/E2E tooling (from the old TRIP-4-test content if present)
TRIP-3-release/SKILL.md (new in v2 — values come from the old TRIP-2)
  1. Start from the new template ($S)
  2. Replace [PROJECT_NAME] with extracted PROJECT_NAME
  3. Replace [VERSION_FILE] with extracted VERSION_FILE
  4. Replace [WEEK_ANCHOR_DATE] with extracted WEEK_ANCHOR_DATE
  5. Replace [MAIN_BRANCH] with the repo's default branch name
  6. Replace the standalone-verification commands with the same extracted lint/typecheck/test commands
  7. Handle tutorial config:
    • If tutorials were disabled: remove the [TUTORIAL_STEP] block
    • If tutorials were enabled: replace the [TUTORIAL_STEP] block with extracted TUTORIAL_CONFIG and renumber subsequent steps
TRIP-review/SKILL.md + checklist.md + cr-template.md (was TRIP-3-review in v1)
  1. SKILL.md: Start from the new template. Replace [PROJECT_NAME].
  2. checklist.md: Start from the new template. Replace the [ADAPT_TO_PROJECT] comment block with the project-specific checklist sections from the extracted REVIEW_CHECKLIST.
    • The new template has generic sections 1-3 (Functional, Code Quality, Architectural) and 4-6 (Error Handling, Security, Performance). The project customization goes between section 3 and 4 (where the comment marker is), and may also modify sections 3-6.
    • If the old checklist had custom sections (numbered 4+), insert them at the [ADAPT_TO_PROJECT] marker and renumber if needed.
    • Preserve the Severity Classification and Approval Gate from the new template unless the project had custom overrides.
  3. cr-template.md: Start from the new template. Update the Checklist section names to match the actual sections in the merged checklist.md.
TRIP-test/SKILL.md (was TRIP-4-test in v1)
  1. Start from the new template ($S)
  2. Replace [PROJECT_NAME] with extracted PROJECT_NAME
  3. Replace [TEST_COMMAND_*] placeholders with extracted TEST_COMMANDS
  4. Replace test structure placeholder with extracted TEST_STRUCTURE
  5. Replace testing priorities placeholder with extracted TESTING_PRIORITIES

New skills: copy from $S. Pure-workflow skills: copy from $S with rsync -a --exclude 'state/' (never delete a skill folder — codex state/ lives inside). Then continue with Phase 4.2 onward, treating the rebuilt files as $MERGED.


Frontmatter and Version Stamp

The YAML frontmatter of every merged SKILL.md (name, description, argument-hint, disable-model-invocation, metadata.trip-version) is pure workflow: take it verbatim from the staged file, never from the installed copy. After the merge, verify every skill reports the staged version:

bash
grep -h "^  trip-version" .claude/skills/*/SKILL.md | sort | uniq -c

Exactly one line must come back, matching the staged version. Any other line means a skill was skipped or merged from the old frontmatter — fix it before reporting completion.


Edge Cases

Installed skills have no metadata.trip-version

Pre-2.8.1 install. Use the structural hints from Phase 1.2 to describe the version, then take the Fallback path. The rebuilt files carry the stamp, so the next upgrade uses the merge.

Installed TRIP-upgrade predates 2.8.4 (or does not exist — v1 installs)

The user copied this skill's folder over .claude/skills/TRIP-upgrade/ by hand to bootstrap. That is expected: run normally. The old skill (if any) used a new-TRIP/ staging folder — if one is present, treat it as a user-supplied $S path and remove it in Phase 5.

Old version has no Codex skills at all

This is the most common upgrade path. The Codex skills are "New" — copy directly. The Codex integration in TRIP-1-plan and TRIP-2-implement comes from the new template and needs no project-specific content except test commands.

Old version has inline checklist but no separate files

The structural migration in Phase 3.1 handles this. Extract the custom checklist content from the old SKILL.md, inject into the new checklist.md.

Old version already has the new structure

Everything categorizes as "Unchanged" or minor updates. The merge is trivial.

Project has extra custom skills not in the new workflow

These are "Removed" in the inventory — warn the user but leave them in place. Never delete skills that exist only in the installed version.

Test commands not available anywhere

If the old version predates the Codex review pre-step and TRIP-4-test doesn't have extractable commands, ask the user:

AskUserQuestion: "The new workflow needs lint/typecheck/test commands for Codex code review. What are the commands for this project?" Options: "Let me provide them" (user types commands) / "Skip for now" (leave placeholders)


Notes for the Agent

  • Read before writing. Read every file you plan to modify. Never write from memory alone.
  • Preserve semantics, not bytes. If the old checklist had 10 custom sections, they all need to survive, even if their numbering changes.
  • New workflow features get project context. When Codex code review is new, the test commands still need to be filled from the project's existing test setup.
  • When in doubt, ask. If you can't confidently extract a customization, show the user the relevant section and ask what to keep.
  • Atomic application. Build all merged content in $MERGED before touching .claude/skills/. If something goes wrong mid-merge, the installed skills are still intact.
  • Never delete a skill folder. Codex state/ directories live inside the skill folders and hold every past review thread. Copy over, never remove-and-copy.
  • Never delete user-created files. If the project has extra files in a skill directory (like project-specific fixtures or notes), leave them alone.

Frequently asked questions

What does the TRIP Upgrade AI skill do?

Upgrade TRIP workflow skills to a newer version while preserving project customizations

Why use TRIP Upgrade on TypingMind?

Because you install it once and use it with any model. TRIP Upgrade is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install TRIP Upgrade in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/PiLastDigit/TRIP-workflow/tree/master/skills/TRIP-upgrade. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use TRIP Upgrade?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with TRIP Upgrade?

As many as you like. As long as a model supports skills, you can use TRIP Upgrade with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the TRIP Upgrade AI skill free?

It is published on GitHub by PiLastDigit. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇