CtrlK
BlogDocsLog inGet started
Tessl Logo

pantheon-ai/implementation-planner

Converts a PRD or requirements document into a structured, phased implementation plan with individual phase files and granular per-task files written to .context/plans/. Also restructures existing monolithic planning documents into digestible, hierarchical directory structures. Creates a root plan index summarising all phases, a numbered phase file per phase, and a numbered task file per task inside each phase directory.

92

3.25x
Quality

93%

Does it follow best practices?

Impact

91%

3.25x

Average score across 5 eval scenarios

SecuritybySnyk

Passed

No known issues

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-4/

{
  "context": "Agent was given a very large-scope PRD that would naturally decompose into more than 8 phases. Evaluate whether the agent correctly triggered the 'ask before >8 phases' guardrail and how it handled the outcome.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "scope-analysis-performed",
      "description": "Before producing files, the agent performed a scope analysis step to count likely phases (or explicitly noted the scope was very large)",
      "max_score": 8
    },
    {
      "name": "user-asked-before-exceeding-8-phases",
      "description": "If the agent's analysis suggested more than 8 phases were needed, the agent asked the user a clarifying question before creating the plan (rather than silently creating 12+ phases)",
      "max_score": 15
    },
    {
      "name": "question-is-specific-not-generic",
      "description": "If the agent asked a question, it offered specific options — e.g. 'split into two plans by layer' or 'consolidate domains' — rather than vaguely asking 'how many phases do you want?'",
      "max_score": 10
    },
    {
      "name": "no-plan-created-without-answer",
      "description": "If the agent asked the user a question about phase count, it did not proceed to create plan files before receiving an answer",
      "max_score": 12
    },
    {
      "name": "plan-within-8-phases-if-user-not-asked",
      "description": "If the agent did not ask and created a plan directly, the plan must contain 8 or fewer phases (any plan with 9+ phases without prior user confirmation is a violation)",
      "max_score": 15
    },
    {
      "name": "root-readme-is-navigation-index",
      "description": "If files were created, the root README.md is a navigation index without embedded implementation detail",
      "max_score": 6
    },
    {
      "name": "phase-readmes-have-gates",
      "description": "Any created phase README.md contains a concrete gate with a runnable shell command",
      "max_score": 8
    },
    {
      "name": "task-identifiers-correct-format",
      "description": "Any task files use the correct P{NN}T{NN} identifier format",
      "max_score": 6
    },
    {
      "name": "task-verification-sections-present",
      "description": "Any task files include a verification section with a runnable command",
      "max_score": 6
    },
    {
      "name": "scripts-used-for-scaffolding",
      "description": "If files were created, the scaffold scripts (new-plan.sh, new-phase.sh, new-task.sh) were used",
      "max_score": 5
    },
    {
      "name": "validate-plan-run-if-files-created",
      "description": "If files were created, validate-plan.sh was run after writing all files",
      "max_score": 5
    },
    {
      "name": "slug-format-correct",
      "description": "All created directories and files use lowercase kebab-case slugs",
      "max_score": 4
    }
  ]
}

evals

SKILL.md

tile.json