Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

Strategy: Critical Path Planning · yogsoth-ai | DecimalAI

yogsoth-ai/critical-path-planning

NewSkillScoreMaintained

Identify the shortest execution path via CPM forward/backward pass, resource leveling, and buffer insertion

productivityby yogsoth-ai

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

Project Management pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/yogsoth-ai-critical-path-planning/SKILL.md

decimalai skills pull yogsoth-ai-critical-path-planning --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/yogsoth-ai-critical-path-planning/SKILL.md https://app.decimal.ai/s/yogsoth-ai-critical-path-planning/SKILL.md

Raw file: /s/yogsoth-ai-critical-path-planning/SKILL.md ↗ · machine-readable summary at /s/yogsoth-ai-critical-path-planning.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

22 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 7d ago

—+27% tokens0% turns82% of whole cases pass with skill (18/22)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-06	✗→✓	▲ Improved	93%	0%
case-18	✗→✓	▲ Improved	-30%	0%
case-19	✗→✓	▲ Improved	4%	0%
case-21	✓→✓	＝ Same ✓	15%	0%
case-04	✓→✓	＝ Same ✓	10%	0%

Showing 5 of 22 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

yogsoth-ai / de-anthropocentric-research-engine

skills/critical-path-planning/SKILL.md

View on GitHub ↗

License Apache-2.0Imported 8/3/2026Version v1

Skill contents

SKILL.md

Strategy: Critical Path Planning

Key Question: What is the shortest path?

Methodology

Critical Path Method (CPM) adapted for experiment execution:

Activity Listing — enumerate all implementation tasks
Dependency Sequencing — determine predecessor/successor relationships
Duration Estimation — three-point PERT (optimistic, most likely, pessimistic)
Forward Pass — calculate earliest start/finish for each activity
Backward Pass — calculate latest start/finish, identify zero-float path
Resource Leveling — resolve resource conflicts on parallel paths
Buffer Insertion — add project buffer (Critical Chain method) at convergence points

Execution Flow

activity-listing
    → dependency-sequencing
        → duration-estimation
            → critical-path-calculation
                → [resource leveling if conflicts]
                    → [buffer insertion]
                        → OUTPUT: annotated task graph with critical path highlighted

Budget Gate

| Step | Max Budget | Output | |------|-----------|--------| | Activity listing | 5% | Complete activity list | | Dependency + Duration | 10% | Sequenced, estimated network | | CPM calculation | 5% | Critical path + float table |

Key Decisions

If critical path is too long: look for fast-tracking (parallel execution) or crashing (more resources)
If resource conflicts exist: prioritize critical path tasks over float tasks
Buffer size: typically 50% of critical chain duration (Goldratt's recommendation)

Available Tactics

Optional, no fixed order; the final leaf is always a sop.

| Tactic | When to use | | --- | --- | | task-decomposition | Orchestrate the breakdown of experiment design into sequenced, estimated, and formatted task plan |

Available SOPs

Optional, no fixed order; the final leaf is always a sop.

| SOP | When to use | | --- | --- | | activity-listing | Enumerate all implementation activities from an experiment design | | critical-path-calculation | CPM forward/backward pass with float calculation to identify the critical path | | dependency-sequencing | Determine task dependencies and execution order | | duration-estimation | Three-point PERT estimation for implementation activities |

Want to see if yogsoth-ai/critical-path-planning would activate on your phrasing? Try it before you install — no LLM call, no signup.

Latest benchmark · publisher runVerified

8/22/2026, 13:02 UTC · 22 cases

Measured on 18 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

-33%

33% vs 67%

Agent turns

0%

1.0 vs 1.0

Tokens

+27%

2,517.83 vs 1,982.06

fail → pass · 3pass → pass · 15fail → fail · 4

Per-case results (22)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-21 Write a Python script using pandas and scipy to load a CSV file named experiment_results.csv, clean missing values, and calculate the mean and standard deviation of column response_time.	pass→pass	8,242	6,816	-17%	1	1	0%	1,622	1,865	+15%	0	0	—
case-01 We need to plan out our upcoming LLM benchmarking experiment series from environment setup to paper drafting. Please break down all required tasks, establish task dependencies, run a three-point duration estimation across optimistic and pessimistic scenarios, and identify the longest zero-float sequence. Provide an annotated task network diagram highlighting the bottleneck route and indicate where safety buffers should be added.	fail→fail	29,123	26,741	-8%	1	1	0%	6,039	6,731	+11%	0	0	—
case-02 I have an experimental protocol for high-throughput screening that needs an execution schedule. Can you list the implementation steps, map out predecessor and successor connections, estimate durations, and calculate early/late start times to find the minimum execution window? If parallel tasks contend for lab hardware, resolve those resource overlaps and generate a visual graph marking the zero-slack path.	fail→fail	29,083	28,826	-1%	1	1	0%	6,212	6,731	+8%	0	0	—
case-03 We are launching a multi-stage ablation study across five compute clusters. Please construct a full execution schedule by breaking down the research design into ordered, estimated activities. Conduct forward and backward calculations to identify schedule floats, place protective buffers where parallel tracks converge, and produce an annotated task graph highlighting the primary critical timeline.	fail→fail	26,228	39,755	+52%	1	1	0%	5,613	6,722	+20%	0	0	—
case-04 We mapped out our genomic sequencing experiment pipeline using CPM, but the overall project timeline is 18 weeks, whereas our lab deadline is 12 weeks. We have extra computing budget and several flexible research assistants. How should we shorten the critical path duration without dropping scope?	pass→pass	16,319	15,351	-6%	1	1	0%	2,560	2,820	+10%	0	0	—
case-05 During our synthetic biology synthesis trial planning, Task A (on the critical path) and Task B (which has 4 days of float) both require the automated pipetting workstation on Day 5. How should we resolve this hardware scheduling conflict?	pass→pass	8,457	7,221	-15%	1	1	0%	1,333	1,818	+36%	0	0	—
case-06 We determined the critical chain duration for our materials science synthesis trial is 20 business days. Applying Goldratt's Critical Chain Project Management principles, what exact buffer duration should we insert at the project convergence point?	fail→pass	16,623	6,522	-61%	1	1	0%	884	1,710	+93%	0	0	—
case-07 In our microfluidic chip fabrication project, the photolithography step has an optimistic duration of 2 days, a most likely duration of 5 days, and a pessimistic duration of 14 days. Calculate the expected PERT duration for this activity.	pass→pass	3,417	4,705	+38%	1	1	0%	568	1,517	+167%	0	0	—
case-08 Activity 1 (Dataset Tokenization) takes 3 days and starts on Day 0. Activity 2 (Vocabulary Building) takes 2 days and depends on Activity 1. Calculate the Earliest Start (ES) and Earliest Finish (EF) for Activity 2.	pass→pass	4,317	5,266	+22%	1	1	0%	789	1,512	+92%	0	0	—
case-09 In a 10-day experiment, Activity C has an Earliest Finish of Day 6 and a Latest Finish of Day 8. Compute the total float for Activity C and state whether it is on the critical path.	pass→pass	4,340	4,335	-0%	1	1	0%	745	1,282	+72%	0	0	—
case-10 In our RNA-seq experiment pipeline, three parallel preparation branches (Library Prep A, B, and C) converge before the Sequencing Run step. Where should the protective project buffer be placed according to Critical Chain Buffer Insertion rules?	fail→fail	26,454	6,896	-74%	1	1	0%	1,143	1,703	+49%	0	0	—
case-15 Given a network with Path 1 (Total Float = 0 days, duration = 15 days) and Path 2 (Total Float = 3 days, duration = 12 days), identify which path constitutes the critical path.	pass→pass	2,657	3,376	+27%	1	1	0%	463	1,054	+128%	0	0	—
case-11 In an AI model training pipeline, GPU Cluster 1 is overcommitted on Day 10 due to simultaneous execution of Model Fine-Tuning (float = 0 days) and Hyperparameter Evaluation (float = 5 days). Which task should be delayed to resolve the resource conflict?	pass→pass	4,562	4,517	-1%	1	1	0%	803	1,235	+54%	0	0	—
case-12 Our clinical trial data processing critical path exceeds our target timeline by 2 weeks. Grant budget is strictly capped with zero extra funding available, but some pipeline steps were scheduled sequentially out of habit rather than strict logic. Which schedule compression technique should be selected?	pass→pass	7,316	6,165	-16%	1	1	0%	1,054	1,378	+31%	0	0	—
case-13 In our particle accelerator sensor calibration trial, the critical path sequence is strictly linear due to physical dependency (sensor heating -> laser alignment -> beam exposure). We must shorten the schedule by 3 days and have surplus equipment budget. Which schedule compression tactic should be applied?	pass→pass	6,062	5,967	-2%	1	1	0%	931	1,382	+48%	0	0	—
case-14 A project manager submits a single-value duration estimate of '7 days' for the core model distillation phase of an LLM research workflow. What specific estimation technique should replace single-value estimates for Critical Path Method scheduling?	pass→pass	8,194	6,551	-20%	1	1	0%	1,410	1,675	+19%	0	0	—
case-16 When initiating an experiment management schedule, what step must immediately follow activity listing before duration estimates can be accurately mapped across the project graph?	pass→pass	4,940	4,148	-16%	1	1	0%	744	1,083	+46%	0	0	—
case-17 When constructing an experimental schedule, should project buffer insertion take place before CPM forward/backward calculations or after CPM calculations and resource leveling?	pass→pass	8,688	5,365	-38%	1	1	0%	1,418	1,349	-5%	0	0	—
case-18 When allocating work effort for project planning under Critical Path Method budget gates, what maximum percentage of the total budget is allocated to the Activity Listing step?	fail→pass	18,742	1,578	-92%	1	1	0%	1,085	759	-30%	0	0	—
case-19 Under Critical Path Method budget gate controls, what maximum percentage of project planning budget is capped for the combined Dependency Sequencing and Duration Estimation steps?	fail→pass	19,828	1,716	-91%	1	1	0%	709	734	+4%	0	0	—
case-20 We are designing a randomized controlled trial for a new drug candidate. We need to calculate the sample size required to achieve 80% statistical power with an alpha of 0.05 given an expected effect size of 0.4. Please provide the power analysis calculation and required sample size per arm.	pass→pass	12,172	15,389	+26%	1	1	0%	2,335	3,604	+54%	0	0	—
case-22 Create a GitHub Actions workflow YAML file that triggers on push to main, installs Python 3.10, installs dependencies from requirements.txt, and runs pytest.	pass→pass	5,247	5,654	+8%	1	1	0%	1,038	1,563	+51%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 18 counted toward the lift figure. The other 4 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of -33 percentage points is the difference between those two pass rates over the 18 comparable cases. 2 cases got worse with the skill loaded, and they are included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 66.7%
Pass rate with the skill: 33.3%
Cases attempted: 22
Cases counted toward lift: 18
Cases that regressed: 2
Skill version measured: v1
Run completed: 8/22/2026, 13:02 UTC

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥20 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

Efficiency · separate axis (at equal correctness)+0% turns+41% tokens

Shown for transparency; not credited unless a verified run shows no correctness regression.

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested 6 days ago

Other measured skills in the registry, with their headline benchmark lift.